每日 AI 资讯 · 持续更新

AI资讯（共 244 篇）AI News (954)

每日精选 AI 行业动态、产品发布、模型更新与趋势洞察，覆盖大模型、智能体、AI 硬件与开源生态。点击卡片阅读完整详情。

最后更新：2026-09-20 · 仅展示最近一周资讯

资讯归档

显示全部资讯

Is the AI industry really ready to slow down?资讯资讯

Is the AI industry really ready to slow down?Is the AI industry really ready to slow down?

We’re seeing a big debate over AI safety and a potential slowdown, as Anthropic CEO Dario Amodei recently published a plWe’re seeing a big debate over AI safety and a potential slowdown, as Anthropic CEO Dario Amodei recently published a pl

📅 2026-09-20 No one is surprised that Nvidia’s Jensen Huang thinks AI fears are overblown.资讯资讯

No one is surprised that Nvidia’s Jensen Huang thinks AI fears are overblown.No one is surprised that Nvidia’s Jensen Huang thinks AI fears are overblown.

The man who may stand to make the most money from the AI boom seems to think he knows better than anyone else, includingThe man who may stand to make the most money from the AI boom seems to think he knows better than anyone else, including

📅 2026-09-20 You too Google! Google Confirms Gemini Breached 3 Companies in AI Security Tests资讯资讯

You too Google! Google Confirms Gemini Breached 3 Companies in AI Security TestsYou too Google! Google Confirms Gemini Breached 3 Companies in AI Security Tests

Google confirmed on Friday, September 18, 2026 that a Gemini model accessed 3 outside companies’ systems. The Wall StreeGoogle confirmed on Friday, September 18, 2026 that a Gemini model accessed 3 outside companies’ systems. The Wall Stree

📅 2026-09-20 Humans, not rogue AI, are still the biggest cybersecurity risk to energy systems资讯资讯

Humans, not rogue AI, are still the biggest cybersecurity risk to energy systemsHumans, not rogue AI, are still the biggest cybersecurity risk to energy systems

Before recent high-profile hacks raised the specter of AI possibly “killing all humans,” our energy systems were alreadyBefore recent high-profile hacks raised the specter of AI possibly “killing all humans,” our energy systems were already

📅 2026-09-20 APUS 开源国内首批Jev跨平台复现：国产模型实现秒级决策资讯资讯

APUS 开源国内首批Jev跨平台复现：国产模型实现秒级决策中文APUS 开源国内首批Jev跨平台复现：国产模型实现秒级决策

APUS 开源国内首批Jev跨平台复现：国产模型实现秒级决策 9月19日，中国人工智能企业APUS旗下 AI 实验室公布了全球最早一批针对Jev的独立开源复现成果 9月19日，中国人工智能企业APUS（麒麟合盛网络技术股份有限公司）旗下 A中文APUS 开源国内首批Jev跨平台复现：国产模型实现秒级决策 9月19日，中国人工智能企业APUS旗下 AI 实验室公布了全球最早一批针对Jev的独立开源复现成果 9月19日，中国人工智能企业APUS（麒麟合盛网络技术股份有限公司）旗下 A

📅 2026-09-20 别把杰作留在实验室，张江见 2026 GDPS 国际具身智能技能大赛·高校赛队招募资讯资讯

别把杰作留在实验室，张江见 2026 GDPS 国际具身智能技能大赛·高校赛队招募中文别把杰作留在实验室，张江见 2026 GDPS 国际具身智能技能大赛·高校赛队招募

具身智能正在从「展示品」变成「生产力」。而这一年，国家把真题摆上了桌面，工信部联合国务院国资委启动《2026年度人形机器人与具身智能实景实训专项行动》，九大重点场景面向全球求解：生产制造、检测分析、维修维护、仓储物流、配送零售、医疗康养、安中文具身智能正在从「展示品」变成「生产力」。而这一年，国家把真题摆上了桌面，工信部联合国务院国资委启动《2026年度人形机器人与具身智能实景实训专项行动》，九大重点场景面向全球求解：生产制造、检测分析、维修维护、仓储物流、配送零售、医疗康养、安

📅 2026-09-20 剪映发布全新AI能力，推进多端智能提效，支持一站式创作资讯资讯

剪映发布全新AI能力，推进多端智能提效，支持一站式创作中文剪映发布全新AI能力，推进多端智能提效，支持一站式创作

0 | 本文作者： Nemo | 2026-09-20 15:07 | 9月20日，剪映举办“AI新创作发布会”，围绕不同创作需求推出剪映Hub、AI创作助手“小映”等多个AI新功能，并升级创作者生态。 从专业创作工作台到移动端AI助手，中文0 | 本文作者： Nemo | 2026-09-20 15:07 | 9月20日，剪映举办“AI新创作发布会”，围绕不同创作需求推出剪映Hub、AI创作助手“小映”等多个AI新功能，并升级创作者生态。 从专业创作工作台到移动端AI助手，

📅 2026-09-20 Alibaba Qwen Team Releases Qwen3.8-LiveTranslate: A Real-Time Interpretation Model That Cuts Average Lag to 2.3 Seconds Across 60 Languages资讯资讯

Alibaba Qwen Team Releases Qwen3.8-LiveTranslate: A Real-Time Interpretation Model That Cuts Average Lag to 2.3 Seconds Across 60 LanguagesAlibaba Qwen Team Releases Qwen3.8-LiveTranslate: A Real-Time Interpretation Model That Cuts Average Lag to 2.3 Seconds Across 60 Languages

Qwen has released Qwen3.8-LiveTranslate, its next-generation real-time simultaneous interpretation model. It listens toQwen has released Qwen3.8-LiveTranslate, its next-generation real-time simultaneous interpretation model. It listens to

📅 2026-09-20 《网络安全人才实战能力报告-AI赋能篇》正式发布，当AI进入业务深水区安全如何跟上资讯资讯

《网络安全人才实战能力报告-AI赋能篇》正式发布，当AI进入业务深水区安全如何跟上中文《网络安全人才实战能力报告-AI赋能篇》正式发布，当AI进入业务深水区安全如何跟上

《网络安全人才实战能力报告-AI赋能篇》正式发布，当AI进入业务深水区安全如何跟上 9月18日，在第一届中国网络空间安全大会上，《网络安全人才实战能力报告—AI赋能篇》正式发布 AI能够独立完成的任务正在迅速变长。人工智能评测机构METR的中文《网络安全人才实战能力报告-AI赋能篇》正式发布，当AI进入业务深水区安全如何跟上 9月18日，在第一届中国网络空间安全大会上，《网络安全人才实战能力报告—AI赋能篇》正式发布 AI能够独立完成的任务正在迅速变长。人工智能评测机构METR的

📅 2026-09-20 沙利文发布2026年智能体市场研究报告：WorkBuddy稳居企业级、个人应用双榜第一资讯资讯

沙利文发布2026年智能体市场研究报告：WorkBuddy稳居企业级、个人应用双榜第一中文沙利文发布2026年智能体市场研究报告：WorkBuddy稳居企业级、个人应用双榜第一

0 | 本文作者： Nemo | 2026-09-20 12:14 | 9月20日，全球权威机构沙利文发布《2026年全球桌面AI智能体市场研究报告》，腾讯WorkBuddy同时位列中国个人端桌面智能体榜单与企业级桌面智能体榜单第一，继续中文0 | 本文作者： Nemo | 2026-09-20 12:14 | 9月20日，全球权威机构沙利文发布《2026年全球桌面AI智能体市场研究报告》，腾讯WorkBuddy同时位列中国个人端桌面智能体榜单与企业级桌面智能体榜单第一，继续

📅 2026-09-20 齐向东：智能体重构软件，安全是必答题资讯资讯

齐向东：智能体重构软件，安全是必答题中文齐向东：智能体重构软件，安全是必答题

0 | 本文作者： 徐咪 | 2026-09-20 11:46 | 9月20日，第28届中国国际软件博览会暨智能体创新应用大会在郑州国际会展中心开幕。全国政协委员、全国工商联副主席、奇安信集团董事长齐向东出席开幕式暨主题峰会，并作《智能体中文0 | 本文作者： 徐咪 | 2026-09-20 11:46 | 9月20日，第28届中国国际软件博览会暨智能体创新应用大会在郑州国际会展中心开幕。全国政协委员、全国工商联副主席、奇安信集团董事长齐向东出席开幕式暨主题峰会，并作《智能体

📅 2026-09-20 阶跃新旗舰 Step 5 Preview 发布，跻身 AA 榜单全球开源前三资讯资讯

阶跃新旗舰 Step 5 Preview 发布，跻身 AA 榜单全球开源前三中文阶跃新旗舰 Step 5 Preview 发布，跻身 AA 榜单全球开源前三

9  月  20  日，阶跃星辰发布新一代旗舰基座模型 Step 5 Preview，重点面向 AI 编程、软件工程、专业知识工作、金融等场景，提升模型在真实世界 Agentic 任务中的综合表现，并进一步拓展中文9  月  20  日，阶跃星辰发布新一代旗舰基座模型 Step 5 Preview，重点面向 AI 编程、软件工程、专业知识工作、金融等场景，提升模型在真实世界 Agentic 任务中的综合表现，并进一步拓展

📅 2026-09-20 OpenClaw Releases 2026.9.5 With Atomic Updates, Plugin Hot Reload, Conversation Sharing, and Expanded GPT Live资讯资讯

OpenClaw Releases 2026.9.5 With Atomic Updates, Plugin Hot Reload, Conversation Sharing, and Expanded GPT LiveOpenClaw Releases 2026.9.5 With Atomic Updates, Plugin Hot Reload, Conversation Sharing, and Expanded GPT Live

OpenClaw is an open-source, MIT-licensed personal AI agent that you run on your own machines. Its Gateway connects modelOpenClaw is an open-source, MIT-licensed personal AI agent that you run on your own machines. Its Gateway connects model

📅 2026-09-19 Google’s Gemini is the latest AI model to hack other companies资讯资讯

Google’s Gemini is the latest AI model to hack other companiesGoogle’s Gemini is the latest AI model to hack other companies

Google’s Gemini accessed the protected systems of three other companies in what The Wall Street Journal reports were theGoogle’s Gemini accessed the protected systems of three other companies in what The Wall Street Journal reports were the

📅 2026-09-19 AI safety conversations have gotten unbelievable资讯资讯

AI safety conversations have gotten unbelievableAI safety conversations have gotten unbelievable

This week two conversations about AI safety went viral that demonstrate just how hard it is to discern AI fact from fictThis week two conversations about AI safety went viral that demonstrate just how hard it is to discern AI fact from fict

📅 2026-09-19 Gemini went rogue, hacked three companies, and Google hid it资讯资讯

Gemini went rogue, hacked three companies, and Google hid itGemini went rogue, hacked three companies, and Google hid it

In May, Gemini broke containment and hacked three different companies, but Google didn’t disclose the incident until theIn May, Gemini broke containment and hacked three different companies, but Google didn’t disclose the incident until the

📅 2026-09-19 TypeSafe AI Releases Jev: A System One Model That Returns Typed, Calibrated Decisions Instead of Text资讯资讯

TypeSafe AI Releases Jev: A System One Model That Returns Typed, Calibrated Decisions Instead of TextTypeSafe AI Releases Jev: A System One Model That Returns Typed, Calibrated Decisions Instead of Text

The ChatGPT moment in 2022 taught AI to talk to people. One of its builders now bets the next moment is AI that talks toThe ChatGPT moment in 2022 taught AI to talk to people. One of its builders now bets the next moment is AI that talks to

📅 2026-09-19 Vals, backed by Andreessen Horowitz, is looking to become the gold standard for AI benchmarking资讯资讯

Vals, backed by Andreessen Horowitz, is looking to become the gold standard for AI benchmarkingVals, backed by Andreessen Horowitz, is looking to become the gold standard for AI benchmarking

Benchmarking has become the industry norm for how AI companies validate their models’ capabilities and, when the metricsBenchmarking has become the industry norm for how AI companies validate their models’ capabilities and, when the metrics

📅 2026-09-19 华为汪涛：华为要打造AI算力底座，只做好一颗芯片远远不够资讯资讯

华为汪涛：华为要打造AI算力底座，只做好一颗芯片远远不够中文华为汪涛：华为要打造AI算力底座，只做好一颗芯片远远不够

华为汪涛：华为要打造AI算力底座，只做好一颗芯片远远不够 CANN跨过拐点，华为补上软件生态 乔不迟 发自 凹非寺 量子位 | 公众号 QbitAI 华为今年全联接大会，最重磅的首先是一组芯片时间表。 昇腾960DT研发进度比原计划提前三个中文华为汪涛：华为要打造AI算力底座，只做好一颗芯片远远不够 CANN跨过拐点，华为补上软件生态 乔不迟 发自 凹非寺 量子位 | 公众号 QbitAI 华为今年全联接大会，最重磅的首先是一组芯片时间表。 昇腾960DT研发进度比原计划提前三个

📅 2026-09-19 AI离“理解万物”还有多远？先拿癌细胞和行星轨道试试水资讯资讯

AI离“理解万物”还有多远？先拿癌细胞和行星轨道试试水中文AI离“理解万物”还有多远？先拿癌细胞和行星轨道试试水

AI离“理解万物”还有多远？先拿癌细胞和行星轨道试试水 同一预测核心，跨七类系统验证 衡宇 发自 凹非寺 量子位 | 公众号 QbitAI 一篇技术报告中居然出现了两个看起来风马牛不相及的研究对象，跟着AI真是见世面。 一个是患者来源的类器中文AI离“理解万物”还有多远？先拿癌细胞和行星轨道试试水 同一预测核心，跨七类系统验证 衡宇 发自 凹非寺 量子位 | 公众号 QbitAI 一篇技术报告中居然出现了两个看起来风马牛不相及的研究对象，跟着AI真是见世面。 一个是患者来源的类器

📅 2026-09-19 Linkup Research Releases SPARSEUP: A 149M-Parameter Open-Source Sparse Embedding Model资讯资讯

Linkup Research Releases SPARSEUP: A 149M-Parameter Open-Source Sparse Embedding ModelLinkup Research Releases SPARSEUP: A 149M-Parameter Open-Source Sparse Embedding Model

Linkup research team releases SPARSEUP, an open-source learned sparse embedding model. The model runs on a 149M-parameteLinkup research team releases SPARSEUP, an open-source learned sparse embedding model. The model runs on a 149M-paramete

📅 2026-09-19 Meta Launches Muse for Mac: A Personal AI Agent That Works Across Your Files, Mail, Messages, Calendar and Notes资讯资讯

Meta Launches Muse for Mac: A Personal AI Agent That Works Across Your Files, Mail, Messages, Calendar and NotesMeta Launches Muse for Mac: A Personal AI Agent That Works Across Your Files, Mail, Messages, Calendar and Notes

Meta has released Muse for Mac, the first version of Muse that can complete things on a user’s computer. The agent worksMeta has released Muse for Mac, the first version of Muse that can complete things on a user’s computer. The agent works

📅 2026-09-19 华为全联接大会2026：360与昇腾AI联合打造解决方案，为AI Agent全面提速资讯资讯

华为全联接大会2026：360与昇腾AI联合打造解决方案，为AI Agent全面提速中文华为全联接大会2026：360与昇腾AI联合打造解决方案，为AI Agent全面提速

0 | 本文作者： Nemo | 2026-09-19 09:58 | 2026年9月17日，第十一届华为全联接大会在上海开幕。大会期间，360集团与昇腾AI联合打造解决方案。以360智能体工厂为核心的智能体基础设施与昇腾AI深度协同，面中文0 | 本文作者： Nemo | 2026-09-19 09:58 | 2026年9月17日，第十一届华为全联接大会在上海开幕。大会期间，360集团与昇腾AI联合打造解决方案。以360智能体工厂为核心的智能体基础设施与昇腾AI深度协同，面

📅 2026-09-19 GGUF vs GPTQ vs AWQ vs EXL2: LLM Model Formats Explained (2026)资讯资讯

GGUF vs GPTQ vs AWQ vs EXL2: LLM Model Formats Explained (2026)GGUF vs GPTQ vs AWQ vs EXL2: LLM Model Formats Explained (2026)

First, separate 2 ideas: containers vs. quantization methods Most confusion comes from mixing 2 layers. A container defiFirst, separate 2 ideas: containers vs. quantization methods Most confusion comes from mixing 2 layers. A container defi

📅 2026-09-19 SpaceXAI Releases Grok Voice Transcribe 2.0: A Speech-to-Text API Claiming 2x Accuracy Over 1.0 at $0.10 per Hour资讯资讯

SpaceXAI Releases Grok Voice Transcribe 2.0: A Speech-to-Text API Claiming 2x Accuracy Over 1.0 at $0.10 per HourSpaceXAI Releases Grok Voice Transcribe 2.0: A Speech-to-Text API Claiming 2x Accuracy Over 1.0 at $0.10 per Hour

SpaceXAI has released Grok Voice Transcribe 2.0, its newest speech-to-text (STT) model. The development team claims it tSpaceXAI has released Grok Voice Transcribe 2.0, its newest speech-to-text (STT) model. The development team claims it t

📅 2026-09-19 A startup that builds other startups raised $100M and is all-in on physical AI资讯资讯

A startup that builds other startups raised $100M and is all-in on physical AIA startup that builds other startups raised $100M and is all-in on physical AI

Four years ago, a startup lab launched that wasn’t quite an incubator, accelerator program, or venture firm. UP.Labs, asFour years ago, a startup lab launched that wasn’t quite an incubator, accelerator program, or venture firm. UP.Labs, as

📅 2026-09-18 Anthropic is operating a lab that conducts biology experiments资讯资讯

Anthropic is operating a lab that conducts biology experimentsAnthropic is operating a lab that conducts biology experiments

Anthropic has a wet biology lab in the Bay Area where it can use its AI models to run physical experiments, it has confiAnthropic has a wet biology lab in the Bay Area where it can use its AI models to run physical experiments, it has confi

📅 2026-09-18 AI hallucination nearly triggers US military operation资讯资讯

AI hallucination nearly triggers US military operationAI hallucination nearly triggers US military operation

Military aircraft were already in the air this spring when U.S. officials made an alarming discovery: The intelligence dMilitary aircraft were already in the air this spring when U.S. officials made an alarming discovery: The intelligence d

📅 2026-09-18 OpenAI and Microsoft knew they were starting a ‘doom loop’ for the web资讯资讯

OpenAI and Microsoft knew they were starting a ‘doom loop’ for the webOpenAI and Microsoft knew they were starting a ‘doom loop’ for the web

Recently unsealed court documents in the New York Times’ case against OpenAI and Microsoft are pretty damning. The compaRecently unsealed court documents in the New York Times’ case against OpenAI and Microsoft are pretty damning. The compa

📅 2026-09-18 Anthropic’s first embedded evaluator is … Accenture?资讯资讯

Anthropic’s first embedded evaluator is … Accenture?Anthropic’s first embedded evaluator is … Accenture?

Dario Amodei’s plans to put third-party safety evaluators inside AI labs are taking shape: Anthropic said that staff froDario Amodei’s plans to put third-party safety evaluators inside AI labs are taking shape: Anthropic said that staff fro

📅 2026-09-18 World model companies are keeping a lot of secrets资讯资讯

World model companies are keeping a lot of secretsWorld model companies are keeping a lot of secrets

This week, I moderated a panel on world models at the All In conference (no relation to the podcast), and it gave me a cThis week, I moderated a panel on world models at the All In conference (no relation to the podcast), and it gave me a c

📅 2026-09-18 A new kind of AI model from a ChatGPT inventor is thrilling developers资讯资讯

A new kind of AI model from a ChatGPT inventor is thrilling developersA new kind of AI model from a ChatGPT inventor is thrilling developers

ChatGPT broke Diogo Almeida’s heart. Almeida was an OpenAI researcher who helped build the chatbot and then invent reinfChatGPT broke Diogo Almeida’s heart. Almeida was an OpenAI researcher who helped build the chatbot and then invent reinf

📅 2026-09-18 Jina AI Releases jina-ocr-v1: A 3.4B MoE Document Parser With Built-In Speculative Decoding for Low-Budget GPUs资讯资讯

Jina AI Releases jina-ocr-v1: A 3.4B MoE Document Parser With Built-In Speculative Decoding for Low-Budget GPUsJina AI Releases jina-ocr-v1: A 3.4B MoE Document Parser With Built-In Speculative Decoding for Low-Budget GPUs

Jina AI, part of Elastic, has released jina-ocr-v1, an end-to-end visual document parser. It takes PDFs, scans, tables,Jina AI, part of Elastic, has released jina-ocr-v1, an end-to-end visual document parser. It takes PDFs, scans, tables,

📅 2026-09-18 无问芯穹与华环电子签署战略合作，共同探索国产异构算力AI基础设施新方向资讯资讯

无问芯穹与华环电子签署战略合作，共同探索国产异构算力AI基础设施新方向中文无问芯穹与华环电子签署战略合作，共同探索国产异构算力AI基础设施新方向

无问芯穹与华环电子签署战略合作，共同探索国产异构算力AI基础设施新方向 协同布局AI基础设施 9月17日，上海无问芯穹智能科技股份有限公司（以下简称“无问芯穹”）与北京华环电子股份有限公司（以下简称“华环电子”）签署战略合作框架协议。双方将中文无问芯穹与华环电子签署战略合作，共同探索国产异构算力AI基础设施新方向 协同布局AI基础设施 9月17日，上海无问芯穹智能科技股份有限公司（以下简称“无问芯穹”）与北京华环电子股份有限公司（以下简称“华环电子”）签署战略合作框架协议。双方将

📅 2026-09-18 达卯科技算电协同2.0平台入选2026国际数字能源展重大成果发布资讯资讯

达卯科技算电协同2.0平台入选2026国际数字能源展重大成果发布中文达卯科技算电协同2.0平台入选2026国际数字能源展重大成果发布

达卯科技算电协同2.0平台入选2026国际数字能源展重大成果发布 成果中唯一聚焦算电协同全链路运营的AI技术产品 9月15日-17日，2026能源绿色发展大会・国际数字能源展在深圳举办。展会首次以“一会一展”深度融合模式打造全球数字能源产业中文达卯科技算电协同2.0平台入选2026国际数字能源展重大成果发布 成果中唯一聚焦算电协同全链路运营的AI技术产品 9月15日-17日，2026能源绿色发展大会・国际数字能源展在深圳举办。展会首次以“一会一展”深度融合模式打造全球数字能源产业

📅 2026-09-18 Google’s new ‘CC’ is an AI agent that helps families run their households资讯资讯

Google’s new ‘CC’ is an AI agent that helps families run their householdsGoogle’s new ‘CC’ is an AI agent that helps families run their households

Google is testing a new product designed to help families coordinate with the support of an AI agent. This week, the seaGoogle is testing a new product designed to help families coordinate with the support of an AI agent. This week, the sea

📅 2026-09-18 Dario Amodei and other AI leaders want to ‘Pace the Frontier’ but…how?资讯资讯

Dario Amodei and other AI leaders want to ‘Pace the Frontier’ but…how?Dario Amodei and other AI leaders want to ‘Pace the Frontier’ but…how?

A week after an Anthropic researcher’s doomsday warning rattled the AI world, the company’s CEO Dario Amodei has outlineA week after an Anthropic researcher’s doomsday warning rattled the AI world, the company’s CEO Dario Amodei has outline

📅 2026-09-18 Automattic’s 33-Hour Coup, and can AI labs police themselves?资讯资讯

Automattic’s 33-Hour Coup, and can AI labs police themselves?Automattic’s 33-Hour Coup, and can AI labs police themselves?

A week after an Anthropic researcher’s doomsday warning rattled the AI world, the company’s CEO Dario Amodei has outlineA week after an Anthropic researcher’s doomsday warning rattled the AI world, the company’s CEO Dario Amodei has outline

📅 2026-09-18 Manus seeks $4B valuation in new $500M fundraise as it resumes independent ops资讯资讯

Manus seeks $4B valuation in new $500M fundraise as it resumes independent opsManus seeks $4B valuation in new $500M fundraise as it resumes independent ops

Chinese AI startup Manus, which earlier this year had to break off a merger with Meta, is in discussions to raise $500 mChinese AI startup Manus, which earlier this year had to break off a merger with Meta, is in discussions to raise $500 m

📅 2026-09-18 Open or closed AI? Nvidia’s Nader Khalil and Sydney Sykes take on one of the decisions shaping next-gen startups at TechCrunch Disrupt 2026资讯资讯

Open or closed AI? Nvidia’s Nader Khalil and Sydney Sykes take on one of the decisions shaping next-gen startups at TechCrunch Disrupt 2026Open or closed AI? Nvidia’s Nader Khalil and Sydney Sykes take on one of the decisions shaping next-gen startups at TechCrunch Disrupt 2026

You have an AI product to build. Do you choose a proprietary frontier model and get moving quickly? Build on an open modYou have an AI product to build. Do you choose a proprietary frontier model and get moving quickly? Build on an open mod

📅 2026-09-18 Meta’s Muse hits Mac, letting the AI take actions on your computer资讯资讯

Meta’s Muse hits Mac, letting the AI take actions on your computerMeta’s Muse hits Mac, letting the AI take actions on your computer

Meta’s new AI assistant app, Muse, is now available on Mac. On the Mac, Muse can interact with your files, messages, calMeta’s new AI assistant app, Muse, is now available on Mac. On the Mac, Muse can interact with your files, messages, cal

📅 2026-09-18 Security researchers used Claude to help them hack into OpenAI资讯资讯

Security researchers used Claude to help them hack into OpenAISecurity researchers used Claude to help them hack into OpenAI

A team of three independent security researchers at Hacktron says it took less than 72 hours for them to hack into OpenAA team of three independent security researchers at Hacktron says it took less than 72 hours for them to hack into OpenA

📅 2026-09-18 PrismML Releases Ternary Bonsai 2 27B: A 5.9 GB Apache 2.0 Model Retaining 98.2% of Qwen3.8 27B Performance资讯资讯

PrismML Releases Ternary Bonsai 2 27B: A 5.9 GB Apache 2.0 Model Retaining 98.2% of Qwen3.8 27B PerformancePrismML Releases Ternary Bonsai 2 27B: A 5.9 GB Apache 2.0 Model Retaining 98.2% of Qwen3.8 27B Performance

PrismML has released Ternary Bonsai 2 27B, a ternary-weight version of Qwen3.8 27B. The language model occupies 5.93 GB,PrismML has released Ternary Bonsai 2 27B, a ternary-weight version of Qwen3.8 27B. The language model occupies 5.93 GB,

📅 2026-09-18 Researchers used Anthropic’s Claude to hack into OpenAI资讯资讯

Researchers used Anthropic’s Claude to hack into OpenAIResearchers used Anthropic’s Claude to hack into OpenAI

In a twist that captures the strange new state of AI security, independent security researchers have used Anthropic’s ClIn a twist that captures the strange new state of AI security, independent security researchers have used Anthropic’s Cl

📅 2026-09-18 AGI最难一战，竟在医院！中国AI登上Science，医生不怕失业还催着上线资讯资讯

AGI最难一战，竟在医院！中国AI登上Science，医生不怕失业还催着上线中文AGI最难一战，竟在医院！中国AI登上Science，医生不怕失业还催着上线

AGI最难一战，竟在医院！中国AI登上Science，医生不怕失业还催着上线 用通用AI去啃最硬的骨头，这条路走得通 克雷西 发自 凹非寺 量子位 | 公众号 QbitAI AI正在横扫一切，但有块硬骨头，迟迟攻不下来，让无数人希望落空。中文AGI最难一战，竟在医院！中国AI登上Science，医生不怕失业还催着上线 用通用AI去啃最硬的骨头，这条路走得通 克雷西 发自 凹非寺 量子位 | 公众号 QbitAI AI正在横扫一切，但有块硬骨头，迟迟攻不下来，让无数人希望落空。

📅 2026-09-18 独家丨皓翊星辰融资千万级美金，厚雪孵化，绿洲、砺思投了资讯资讯

独家丨皓翊星辰融资千万级美金，厚雪孵化，绿洲、砺思投了中文独家丨皓翊星辰融资千万级美金，厚雪孵化，绿洲、砺思投了

网球发球机一年量产交付，篮球训练机器人众筹已破百万。 作者丨陈玘凡 雷峰网·鲸犀独家获悉，皓翊星辰（LUMISTAR）已完成绿洲资本、砺思资本（Monolith）投资的多轮融资，累计融资额达千万级美元。目前新一轮融资正在洽谈中。资金将用于产中文网球发球机一年量产交付，篮球训练机器人众筹已破百万。 作者丨陈玘凡 雷峰网·鲸犀独家获悉，皓翊星辰（LUMISTAR）已完成绿洲资本、砺思资本（Monolith）投资的多轮融资，累计融资额达千万级美元。目前新一轮融资正在洽谈中。资金将用于产

📅 2026-09-18 仅用3天！千问办公协助国家天文台科研团队打造科研级望远镜仿真系统资讯资讯

仅用3天！千问办公协助国家天文台科研团队打造科研级望远镜仿真系统中文仅用3天！千问办公协助国家天文台科研团队打造科研级望远镜仿真系统

日前，千问办公协助国家天文台科研团队打造大口径科研级望远镜数字化仿真系统，耗时仅3天，成本不到千元。依托对真实望远镜组件和观测条件的模拟，Agent可在仿真环境中驱动望远镜进行远程态势感知与观测流程仿真。未来，仿真系统产生的数据，还将用于训中文日前，千问办公协助国家天文台科研团队打造大口径科研级望远镜数字化仿真系统，耗时仅3天，成本不到千元。依托对真实望远镜组件和观测条件的模拟，Agent可在仿真环境中驱动望远镜进行远程态势感知与观测流程仿真。未来，仿真系统产生的数据，还将用于训

📅 2026-09-18 独家丨OpenAI的一款无屏硬件，正在深圳试制资讯资讯

独家丨OpenAI的一款无屏硬件，正在深圳试制中文独家丨OpenAI的一款无屏硬件，正在深圳试制

0 雷峰网(公众号：雷峰网)·鲸犀获悉，OpenAI近日正在深圳秘密试制一款无屏AI硬件产品。有业内人士称，其保密程度极高，工程师来深圳出差连手机都没带。该产品计划在2027年上市。这会是一款什么形态的产品？还有哪些最新进展？欢迎添加作者微中文0 雷峰网(公众号：雷峰网)·鲸犀获悉，OpenAI近日正在深圳秘密试制一款无屏AI硬件产品。有业内人士称，其保密程度极高，工程师来深圳出差连手机都没带。该产品计划在2027年上市。这会是一款什么形态的产品？还有哪些最新进展？欢迎添加作者微

📅 2026-09-18 Best Open-Source Agent Harnesses for Local LLMs in 2026资讯资讯

Best Open-Source Agent Harnesses for Local LLMs in 2026Best Open-Source Agent Harnesses for Local LLMs in 2026

An agent is a model and a harness. The harness runs tools, holds state, manages permissions, and feeds context back to tAn agent is a model and a harness. The harness runs tools, holds state, manages permissions, and feeds context back to t

📅 2026-09-18 Alibaba Qwen Releases Qwen3.8-Omni-Flash: A 1M-Context Omni-Modal Model Built Around Agentic Audio-Video Understanding and Tool Use资讯资讯

Alibaba Qwen Releases Qwen3.8-Omni-Flash: A 1M-Context Omni-Modal Model Built Around Agentic Audio-Video Understanding and Tool UseAlibaba Qwen Releases Qwen3.8-Omni-Flash: A 1M-Context Omni-Modal Model Built Around Agentic Audio-Video Understanding and Tool Use

Alibaba’s Qwen team has released Qwen3.8-Omni-Flash. They called it its first omni-modal model built around agentic capaAlibaba’s Qwen team has released Qwen3.8-Omni-Flash. They called it its first omni-modal model built around agentic capa

📅 2026-09-18 Salesforce Agentforce: Bridging the Enterprise AI Gap from ‘Vibe Coding’ to Battle-Tested Orchestration资讯资讯

Salesforce Agentforce: Bridging the Enterprise AI Gap from ‘Vibe Coding’ to Battle-Tested OrchestrationSalesforce Agentforce: Bridging the Enterprise AI Gap from ‘Vibe Coding’ to Battle-Tested Orchestration

Spinning up a flashy prototype with an LLM has never been easier, but in production, building an agent is just the openiSpinning up a flashy prototype with an LLM has never been easier, but in production, building an agent is just the openi

📅 2026-09-18 Regularized Emphatic Temporal-Difference Learning: Stability under Constant Stepsizes资讯资讯

Regularized Emphatic Temporal-Difference Learning: Stability under Constant StepsizesRegularized Emphatic Temporal-Difference Learning: Stability under Constant Stepsizes

Computer Science > Artificial Intelligence [Submitted on 14 Sep 2026] Title:Regularized Emphatic Temporal-DifferenComputer Science > Artificial Intelligence [Submitted on 14 Sep 2026] Title:Regularized Emphatic Temporal-Differen

📅 2026-09-18 BioPhys-Bridge: A Benchmark for Interdisciplinary Scientific Reasoning in Physics-Grounded Biological Research资讯资讯

BioPhys-Bridge: A Benchmark for Interdisciplinary Scientific Reasoning in Physics-Grounded Biological ResearchBioPhys-Bridge: A Benchmark for Interdisciplinary Scientific Reasoning in Physics-Grounded Biological Research

Computer Science > Artificial Intelligence [Submitted on 15 Sep 2026] Title:BioPhys-Bridge: A Benchmark for InterdComputer Science > Artificial Intelligence [Submitted on 15 Sep 2026] Title:BioPhys-Bridge: A Benchmark for Interd

📅 2026-09-18 What Do We Expect from LLMs? Mapping the Design of LLM Benchmarks资讯资讯

What Do We Expect from LLMs? Mapping the Design of LLM BenchmarksWhat Do We Expect from LLMs? Mapping the Design of LLM Benchmarks

Computer Science > Artificial Intelligence [Submitted on 15 Sep 2026] Title:What Do We Expect from LLMs? Mapping tComputer Science > Artificial Intelligence [Submitted on 15 Sep 2026] Title:What Do We Expect from LLMs? Mapping t

📅 2026-09-18 Position: It is Time to Virtualize Foundation Models with a Self-evolving Operating System Layer资讯资讯

Position: It is Time to Virtualize Foundation Models with a Self-evolving Operating System LayerPosition: It is Time to Virtualize Foundation Models with a Self-evolving Operating System Layer

Computer Science > Artificial Intelligence [Submitted on 16 Sep 2026] Title:Position: It is Time to Virtualize FouComputer Science > Artificial Intelligence [Submitted on 16 Sep 2026] Title:Position: It is Time to Virtualize Fou

📅 2026-09-18 What Do Current Systematic Generalization Tasks Miss? A Reasoning-Centered Analysis资讯资讯

What Do Current Systematic Generalization Tasks Miss? A Reasoning-Centered AnalysisWhat Do Current Systematic Generalization Tasks Miss? A Reasoning-Centered Analysis

Computer Science > Artificial Intelligence [Submitted on 16 Sep 2026] Title:What Do Current Systematic GeneralizatComputer Science > Artificial Intelligence [Submitted on 16 Sep 2026] Title:What Do Current Systematic Generalizat

📅 2026-09-18 Characterizing Web Search by Conversational LLM Agents: From Search Decisions and Strategies to Results and Responses资讯资讯

Characterizing Web Search by Conversational LLM Agents: From Search Decisions and Strategies to Results and ResponsesCharacterizing Web Search by Conversational LLM Agents: From Search Decisions and Strategies to Results and Responses

Computer Science > Artificial Intelligence [Submitted on 16 Sep 2026] Title:Characterizing Web Search by ConversatComputer Science > Artificial Intelligence [Submitted on 16 Sep 2026] Title:Characterizing Web Search by Conversat

📅 2026-09-18 Do AI Agents Understand Computer Architecture?资讯资讯

Do AI Agents Understand Computer Architecture?Do AI Agents Understand Computer Architecture?

Computer Science > Artificial Intelligence [Submitted on 16 Sep 2026] Title:Do AI Agents Understand Computer ArchiComputer Science > Artificial Intelligence [Submitted on 16 Sep 2026] Title:Do AI Agents Understand Computer Archi

📅 2026-09-18 MAGS: Multi-agent Auto-formalization Guarantees Safety for Agentic Outputs资讯资讯

MAGS: Multi-agent Auto-formalization Guarantees Safety for Agentic OutputsMAGS: Multi-agent Auto-formalization Guarantees Safety for Agentic Outputs

Computer Science > Artificial Intelligence [Submitted on 16 Sep 2026] Title:MAGS: Multi-agent Auto-formalization GComputer Science > Artificial Intelligence [Submitted on 16 Sep 2026] Title:MAGS: Multi-agent Auto-formalization G

📅 2026-09-18 Closed-World Resolution Against Tool Hallucination in LLM Agents资讯资讯

Closed-World Resolution Against Tool Hallucination in LLM AgentsClosed-World Resolution Against Tool Hallucination in LLM Agents

Computer Science > Artificial Intelligence [Submitted on 16 Sep 2026] Title:Closed-World Resolution Against Tool HComputer Science > Artificial Intelligence [Submitted on 16 Sep 2026] Title:Closed-World Resolution Against Tool H

📅 2026-09-18 The syntax and semantics of goals资讯资讯

The syntax and semantics of goalsThe syntax and semantics of goals

Computer Science > Artificial Intelligence [Submitted on 16 Sep 2026] Title:The syntax and semantics of goals ViewComputer Science > Artificial Intelligence [Submitted on 16 Sep 2026] Title:The syntax and semantics of goals View

📅 2026-09-18 Compositional Reasoning in Language Models under Reinforcement Learning Post-Training资讯资讯

Compositional Reasoning in Language Models under Reinforcement Learning Post-TrainingCompositional Reasoning in Language Models under Reinforcement Learning Post-Training

Computer Science > Artificial Intelligence [Submitted on 16 Sep 2026] Title:Compositional Reasoning in Language MoComputer Science > Artificial Intelligence [Submitted on 16 Sep 2026] Title:Compositional Reasoning in Language Mo

📅 2026-09-18 Safety Beyond the Interface: Detecting Harm via Latent States in Large Language Models资讯资讯

Safety Beyond the Interface: Detecting Harm via Latent States in Large Language ModelsSafety Beyond the Interface: Detecting Harm via Latent States in Large Language Models

Computer Science > Artificial Intelligence [Submitted on 16 Sep 2026] Title:Safety Beyond the Interface: DetectingComputer Science > Artificial Intelligence [Submitted on 16 Sep 2026] Title:Safety Beyond the Interface: Detecting

📅 2026-09-18 QVAC Genesis III: A Large-Scale, High-Quality Open Synthetic STEM Corpus for Efficient Language Model Pre-Training资讯资讯

QVAC Genesis III: A Large-Scale, High-Quality Open Synthetic STEM Corpus for Efficient Language Model Pre-TrainingQVAC Genesis III: A Large-Scale, High-Quality Open Synthetic STEM Corpus for Efficient Language Model Pre-Training

Computer Science > Artificial Intelligence [Submitted on 17 Sep 2026] Title:QVAC Genesis III: A Large-Scale, High-Computer Science > Artificial Intelligence [Submitted on 17 Sep 2026] Title:QVAC Genesis III: A Large-Scale, High-

📅 2026-09-18 LLM-as-an-Improver: Turning Verification into Better Candidates资讯资讯

LLM-as-an-Improver: Turning Verification into Better CandidatesLLM-as-an-Improver: Turning Verification into Better Candidates

Computer Science > Artificial Intelligence [Submitted on 17 Sep 2026] Title:LLM-as-an-Improver: Turning VerificatiComputer Science > Artificial Intelligence [Submitted on 17 Sep 2026] Title:LLM-as-an-Improver: Turning Verificati

📅 2026-09-18 An Architecture for Long-Horizon Agents: Levels, Ticks and Cascaded Intelligence资讯资讯

An Architecture for Long-Horizon Agents: Levels, Ticks and Cascaded IntelligenceAn Architecture for Long-Horizon Agents: Levels, Ticks and Cascaded Intelligence

Computer Science > Artificial Intelligence [Submitted on 17 Sep 2026] Title:An Architecture for Long-Horizon AgentComputer Science > Artificial Intelligence [Submitted on 17 Sep 2026] Title:An Architecture for Long-Horizon Agent

📅 2026-09-18 Subliminal Prompting Beyond Static Geometry: Causal Depth and Multi-Token Confounds资讯资讯

Subliminal Prompting Beyond Static Geometry: Causal Depth and Multi-Token ConfoundsSubliminal Prompting Beyond Static Geometry: Causal Depth and Multi-Token Confounds

Computer Science > Computation and Language [Submitted on 20 Jul 2026] Title:Subliminal Prompting Beyond Static GeComputer Science > Computation and Language [Submitted on 20 Jul 2026] Title:Subliminal Prompting Beyond Static Ge

📅 2026-09-18 Sampling Reveals Style: Unsupervised, Training-Free Discovery of Prompt-Conditional Stylistic Axes in LLM Activations资讯资讯

Sampling Reveals Style: Unsupervised, Training-Free Discovery of Prompt-Conditional Stylistic Axes in LLM ActivationsSampling Reveals Style: Unsupervised, Training-Free Discovery of Prompt-Conditional Stylistic Axes in LLM Activations

Computer Science > Computation and Language [Submitted on 20 Jul 2026] Title:Sampling Reveals Style: Unsupervised,Computer Science > Computation and Language [Submitted on 20 Jul 2026] Title:Sampling Reveals Style: Unsupervised,

📅 2026-09-18 FakeSpotter: A content and strategy agnostic Viral Misinformation Detection Tool资讯资讯

FakeSpotter: A content and strategy agnostic Viral Misinformation Detection ToolFakeSpotter: A content and strategy agnostic Viral Misinformation Detection Tool

Computer Science > Computation and Language [Submitted on 20 Jul 2026] Title:FakeSpotter: A content and strategy aComputer Science > Computation and Language [Submitted on 20 Jul 2026] Title:FakeSpotter: A content and strategy a

📅 2026-09-18 Towards Proactive Detection of User-Side Implicit Conflicts in Human-LLM Dialogue资讯资讯

Towards Proactive Detection of User-Side Implicit Conflicts in Human-LLM DialogueTowards Proactive Detection of User-Side Implicit Conflicts in Human-LLM Dialogue

Computer Science > Computation and Language [Submitted on 22 Jul 2026] Title:Towards Proactive Detection of User-SComputer Science > Computation and Language [Submitted on 22 Jul 2026] Title:Towards Proactive Detection of User-S

📅 2026-09-18 Reflective Recovery: A Self-Supervised Method for Reasoning by Learning from Mistakes资讯资讯

Reflective Recovery: A Self-Supervised Method for Reasoning by Learning from MistakesReflective Recovery: A Self-Supervised Method for Reasoning by Learning from Mistakes

Computer Science > Computation and Language [Submitted on 24 Jul 2026] Title:Reflective Recovery: A Self-SuperviseComputer Science > Computation and Language [Submitted on 24 Jul 2026] Title:Reflective Recovery: A Self-Supervise

📅 2026-09-18 VisKG-LM: Compiling Knowledge Graphs into Visual Memory for Multiple-Choice Question Answering资讯资讯

VisKG-LM: Compiling Knowledge Graphs into Visual Memory for Multiple-Choice Question AnsweringVisKG-LM: Compiling Knowledge Graphs into Visual Memory for Multiple-Choice Question Answering

Computer Science > Computation and Language [Submitted on 29 Jul 2026] Title:VisKG-LM: Compiling Knowledge GraphsComputer Science > Computation and Language [Submitted on 29 Jul 2026] Title:VisKG-LM: Compiling Knowledge Graphs

📅 2026-09-18 Advantage Scale Calibration Imbalance in Group-Relative Optimization under Low-Variance Rewards: Diagnosis and Bounded Recovery资讯资讯

Advantage Scale Calibration Imbalance in Group-Relative Optimization under Low-Variance Rewards: Diagnosis and Bounded RecoveryAdvantage Scale Calibration Imbalance in Group-Relative Optimization under Low-Variance Rewards: Diagnosis and Bounded Recovery

Computer Science > Computation and Language [Submitted on 27 Aug 2026] Title:Advantage Scale Calibration ImbalanceComputer Science > Computation and Language [Submitted on 27 Aug 2026] Title:Advantage Scale Calibration Imbalance

📅 2026-09-18 To Memories and Beyond: From Remembering to Knowing You across Long-Term Multimodal Personal Archives资讯资讯

To Memories and Beyond: From Remembering to Knowing You across Long-Term Multimodal Personal ArchivesTo Memories and Beyond: From Remembering to Knowing You across Long-Term Multimodal Personal Archives

Computer Science > Computation and Language [Submitted on 11 Sep 2026] Title:To Memories and Beyond: From RememberComputer Science > Computation and Language [Submitted on 11 Sep 2026] Title:To Memories and Beyond: From Remember

📅 2026-09-18 Why Pretraining Fails to Share Cross-Lingual Knowledge资讯资讯

Why Pretraining Fails to Share Cross-Lingual KnowledgeWhy Pretraining Fails to Share Cross-Lingual Knowledge

Computer Science > Computation and Language [Submitted on 16 Sep 2026] Title:Why Pretraining Fails to Share Cross-Computer Science > Computation and Language [Submitted on 16 Sep 2026] Title:Why Pretraining Fails to Share Cross-

📅 2026-09-18 A frontend-backend architecture for tool calls in full-duplex speech models资讯资讯

A frontend-backend architecture for tool calls in full-duplex speech modelsA frontend-backend architecture for tool calls in full-duplex speech models

Computer Science > Computation and Language [Submitted on 16 Sep 2026] Title:A frontend-backend architecture for tComputer Science > Computation and Language [Submitted on 16 Sep 2026] Title:A frontend-backend architecture for t

📅 2026-09-18 Crusoe raises $3.9B to build massive data centers and small modular ‘AI factories’资讯资讯

Crusoe raises $3.9B to build massive data centers and small modular ‘AI factories’Crusoe raises $3.9B to build massive data centers and small modular ‘AI factories’

Data center developer Crusoe said Thursday it raised $3.9 billion in a Series F round that pushes its valuation to $30.9Data center developer Crusoe said Thursday it raised $3.9 billion in a Series F round that pushes its valuation to $30.9

📅 2026-09-17 Google DeepMind launches institute to widen the AGI debate资讯资讯

Google DeepMind launches institute to widen the AGI debateGoogle DeepMind launches institute to widen the AGI debate

Google and Google DeepMind researchers launched the DeepMind Institute on Wednesday to advance the conversation around aGoogle and Google DeepMind researchers launched the DeepMind Institute on Wednesday to advance the conversation around a

📅 2026-09-17 PrismML hopes its tiny LLM could change how we all use AI资讯资讯

PrismML hopes its tiny LLM could change how we all use AIPrismML hopes its tiny LLM could change how we all use AI

If AI lab PrismML isn’t on your radar yet, it should be — not because it’s raised gobs of money (it hasn’t yet, just a $If AI lab PrismML isn’t on your radar yet, it should be — not because it’s raised gobs of money (it hasn’t yet, just a $

📅 2026-09-17 The FAA’s plan to fix air traffic? $875 million worth of AI资讯资讯

The FAA’s plan to fix air traffic? $875 million worth of AIThe FAA’s plan to fix air traffic? $875 million worth of AI

The Federal Aviation Administration has been struggling to manage an air traffic control shortage throughout the countryThe Federal Aviation Administration has been struggling to manage an air traffic control shortage throughout the country

📅 2026-09-17 OpenAI caught its models leaving notes to successors to hide bad behavior资讯资讯

OpenAI caught its models leaving notes to successors to hide bad behaviorOpenAI caught its models leaving notes to successors to hide bad behavior

OpenAI caught something unusual while training its latest model, GPT-5.6 Sol: It began leaving instructions for future vOpenAI caught something unusual while training its latest model, GPT-5.6 Sol: It began leaving instructions for future v

📅 2026-09-17 UN turns to Google to make its global data ready for AI agents资讯资讯

UN turns to Google to make its global data ready for AI agentsUN turns to Google to make its global data ready for AI agents

The United Nations on Thursday announced that it is working with Google to make its vast collection of global statisticsThe United Nations on Thursday announced that it is working with Google to make its vast collection of global statistics

📅 2026-09-17 Microsoft exec called AI scraping ‘the largest theft of labor in human history,’ new unredacted filings reveal资讯资讯

Microsoft exec called AI scraping ‘the largest theft of labor in human history,’ new unredacted filings revealMicrosoft exec called AI scraping ‘the largest theft of labor in human history,’ new unredacted filings reveal

New unredacted information in the copyright lawsuit The New York Times brought against OpenAI and Microsoft three yearsNew unredacted information in the copyright lawsuit The New York Times brought against OpenAI and Microsoft three years

📅 2026-09-17 The AI Superintelligence Slowdown资讯资讯

The AI Superintelligence SlowdownThe AI Superintelligence Slowdown

Remember when tech leaders would tell their employees to “move fast and break things”? It seemed that would be the way oRemember when tech leaders would tell their employees to “move fast and break things”? It seemed that would be the way o

📅 2026-09-17 Claude Code relaunches Projects to manage multiple AI agents in the cloud资讯资讯

Claude Code relaunches Projects to manage multiple AI agents in the cloudClaude Code relaunches Projects to manage multiple AI agents in the cloud

The revamped projects feature in Claude Code allows users to run multiple agents under the same roof, with a shared memoThe revamped projects feature in Claude Code allows users to run multiple agents under the same roof, with a shared memo

📅 2026-09-17 Microsoft Open-Sources TauGrid: A Kubernetes-Native Stack for GPU AI Workloads资讯资讯

Microsoft Open-Sources TauGrid: A Kubernetes-Native Stack for GPU AI WorkloadsMicrosoft Open-Sources TauGrid: A Kubernetes-Native Stack for GPU AI Workloads

Platform teams running AI on Kubernetes rarely run one thing. They run a queueing system, a distributed runtime, GPU nodPlatform teams running AI on Kubernetes rarely run one thing. They run a queueing system, a distributed runtime, GPU nod

📅 2026-09-17 Anthropic Launches Claude Code Projects in Beta: Parallel Cloud Sessions That Keep Running After You Close Your Laptop资讯资讯

Anthropic Launches Claude Code Projects in Beta: Parallel Cloud Sessions That Keep Running After You Close Your LaptopAnthropic Launches Claude Code Projects in Beta: Parallel Cloud Sessions That Keep Running After You Close Your Laptop

Anthropic redesigned Projects in Claude Code. The old project was a folder: some files plus one chat. The new one is a sAnthropic redesigned Projects in Claude Code. The old project was a folder: some files plus one chat. The new one is a s

📅 2026-09-17 华为开创AI时代计算架构：让百万处理器成为一台计算机资讯资讯

华为开创AI时代计算架构：让百万处理器成为一台计算机中文华为开创AI时代计算架构：让百万处理器成为一台计算机

0 | 本文作者： 於 | 2026-09-17 22:25 | 关键信息： • Peerium计算架构突破了图灵范式，提出了Nested BSP，突破了冯·诺依曼单机架构，颠覆了长久以来的主从架构，实现了百万级处理器真正成为一台更大的计中文0 | 本文作者： 於 | 2026-09-17 22:25 | 关键信息： • Peerium计算架构突破了图灵范式，提出了Nested BSP，突破了冯·诺依曼单机架构，颠覆了长久以来的主从架构，实现了百万级处理器真正成为一台更大的计

📅 2026-09-17 Base Labs launches an open-weight AI safety partnership with Hugging Face and Goodfire资讯资讯

Base Labs launches an open-weight AI safety partnership with Hugging Face and GoodfireBase Labs launches an open-weight AI safety partnership with Hugging Face and Goodfire

Baseten launched a new safety infrastructure standard alongside its Base Labs research arm on Wednesday, partnering withBaseten launched a new safety infrastructure standard alongside its Base Labs research arm on Wednesday, partnering with

📅 2026-09-17 Pinterest teases a new ‘Restyle’ feature that lets you redesign your room with AI资讯资讯

Pinterest teases a new ‘Restyle’ feature that lets you redesign your room with AIPinterest teases a new ‘Restyle’ feature that lets you redesign your room with AI

Pinterest is turning to AI to help consumers move from searching for product inspiration to actually being able to imagiPinterest is turning to AI to help consumers move from searching for product inspiration to actually being able to imagi

📅 2026-09-17 Claude双入口合并，原生Office上线！硅谷AI办公大战也开始了资讯资讯

Claude双入口合并，原生Office上线！硅谷AI办公大战也开始了中文Claude双入口合并，原生Office上线！硅谷AI办公大战也开始了

Claude双入口合并，原生Office上线！硅谷AI办公大战也开始了 文档和PPT都能做了 听雨 发自 凹非寺 量子位 | 公众号QbitAI 好家伙，现在AI办公大战已经卷到大洋彼岸了…… 就在刚刚，Anthropic宣布Cowork与中文Claude双入口合并，原生Office上线！硅谷AI办公大战也开始了 文档和PPT都能做了 听雨 发自 凹非寺 量子位 | 公众号QbitAI 好家伙，现在AI办公大战已经卷到大洋彼岸了…… 就在刚刚，Anthropic宣布Cowork与

📅 2026-09-17 安芯启程｜安芯724北京基地正式投产资讯资讯

安芯启程｜安芯724北京基地正式投产中文安芯启程｜安芯724北京基地正式投产

0 | 本文作者： Nemo | 2026-09-17 17:44 | 2026年9月16日，柒贰肆安芯技术服务有限公司（简称“安芯724”）北京基地投产活动暨品牌发布会在北京市昌平区未来星科能源谷智造产业园举行。安芯724北京基地正式投中文0 | 本文作者： Nemo | 2026-09-17 17:44 | 2026年9月16日，柒贰肆安芯技术服务有限公司（简称“安芯724”）北京基地投产活动暨品牌发布会在北京市昌平区未来星科能源谷智造产业园举行。安芯724北京基地正式投

📅 2026-09-17 Huawei plans Q1 2027 launch of new AI chip as it takes on Nvidia资讯资讯

Huawei plans Q1 2027 launch of new AI chip as it takes on NvidiaHuawei plans Q1 2027 launch of new AI chip as it takes on Nvidia

Huawei is moving up the launch of its next-generation Ascend 960DT AI chip to the first quarter of 2027, as it races toHuawei is moving up the launch of its next-generation Ascend 960DT AI chip to the first quarter of 2027, as it races to

📅 2026-09-17 Rival AI agents, Instinct and Meta’s Muse, both add the ability to make calls资讯资讯

Rival AI agents, Instinct and Meta’s Muse, both add the ability to make callsRival AI agents, Instinct and Meta’s Muse, both add the ability to make calls

Text-based AI assistants are a hot market, with Instinct, Wajo, Town, Ollie, and Meta’s Muse agent, which launched lastText-based AI assistants are a hot market, with Instinct, Wajo, Town, Ollie, and Meta’s Muse agent, which launched last

📅 2026-09-17 Google, Nvidia, and Anthropic want Emerald AI to find space on the grid for more data centers资讯资讯

Google, Nvidia, and Anthropic want Emerald AI to find space on the grid for more data centersGoogle, Nvidia, and Anthropic want Emerald AI to find space on the grid for more data centers

Grid software unicorn Emerald AI has formed a coalition with Google and Nvidia that will use the startup’s technology toGrid software unicorn Emerald AI has formed a coalition with Google and Nvidia that will use the startup’s technology to

📅 2026-09-17 Microsoft AI CEO says AI threats are real, and Anthropic is making it worse资讯资讯

Microsoft AI CEO says AI threats are real, and Anthropic is making it worseMicrosoft AI CEO says AI threats are real, and Anthropic is making it worse

Today, I’m talking with Mustafa Suleyman, the CEO of Microsoft AI. As you’re no doubt aware, the biggest story in tech rToday, I’m talking with Mustafa Suleyman, the CEO of Microsoft AI. As you’re no doubt aware, the biggest story in tech r

📅 2026-09-17 蚂蚁集团全员接入千问办公，打造大型企业办公标杆资讯资讯

蚂蚁集团全员接入千问办公，打造大型企业办公标杆中文蚂蚁集团全员接入千问办公，打造大型企业办公标杆

9月17日，记者获悉，蚂蚁集团正式接入千问办公，作为全公司的智能办公Agent底座，供全员使用，打造大型企业办公的标杆。  图示 蚂蚁集团和千问办公达成合作 蚂蚁集团是全球领先的数字支付与科技服务平台，业务涵盖中文9月17日，记者获悉，蚂蚁集团正式接入千问办公，作为全公司的智能办公Agent底座，供全员使用，打造大型企业办公的标杆。  图示 蚂蚁集团和千问办公达成合作 蚂蚁集团是全球领先的数字支付与科技服务平台，业务涵盖

📅 2026-09-17 独家丨前腾讯混元预训练负责人姚星丞，加入Thinking Machines Lab资讯资讯

独家丨前腾讯混元预训练负责人姚星丞，加入Thinking Machines Lab中文独家丨前腾讯混元预训练负责人姚星丞，加入Thinking Machines Lab

0 雷峰网独家获悉，腾讯混元大语言模型预训练负责人姚星丞，近日已加入北美创业公司Thinking Machines Lab。 据知情人士透露，Thinking Machines Lab此次给姚星丞开出的薪资，明显高于其在腾讯的收入。而姚星丞中文0 雷峰网独家获悉，腾讯混元大语言模型预训练负责人姚星丞，近日已加入北美创业公司Thinking Machines Lab。 据知情人士透露，Thinking Machines Lab此次给姚星丞开出的薪资，明显高于其在腾讯的收入。而姚星丞

📅 2026-09-17 豆包座舱助手发布，首款合作车型即将开启预售资讯资讯

豆包座舱助手发布，首款合作车型即将开启预售中文豆包座舱助手发布，首款合作车型即将开启预售

0 | 本文作者： Nemo | 2026-09-17 15:03 | 9月17日，豆包与火山引擎联合打造的AI原生车载助手——豆包座舱助手正式发布。 豆包座舱助手致力于成为聪明懂你、交互自然有温度、安全可靠的座舱 AI 伙伴，带来更加方中文0 | 本文作者： Nemo | 2026-09-17 15:03 | 9月17日，豆包与火山引擎联合打造的AI原生车载助手——豆包座舱助手正式发布。 豆包座舱助手致力于成为聪明懂你、交互自然有温度、安全可靠的座舱 AI 伙伴，带来更加方

📅 2026-09-17 OpenAI Releases a Model Misalignment Disclosure Framework With 3 Review Tracks and 6 Incident Reports From RL Training资讯资讯

OpenAI Releases a Model Misalignment Disclosure Framework With 3 Review Tracks and 6 Incident Reports From RL TrainingOpenAI Releases a Model Misalignment Disclosure Framework With 3 Review Tracks and 6 Incident Reports From RL Training

OpenAI has released a new framework for tracking, investigating, and disclosing misalignment in its own models. The OpenOpenAI has released a new framework for tracking, investigating, and disclosing misalignment in its own models. The Open

📅 2026-09-17 Google Research Introduces Retrieve-for-Train (R4T): An RL-Compiled Diffusion Retriever for 12× to 20× Faster Query Fan-Out资讯资讯

Google Research Introduces Retrieve-for-Train (R4T): An RL-Compiled Diffusion Retriever for 12× to 20× Faster Query Fan-OutGoogle Research Introduces Retrieve-for-Train (R4T): An RL-Compiled Diffusion Retriever for 12× to 20× Faster Query Fan-Out

Search and recommendation systems increasingly need to return a set of results, not one best match. A query like ‘campinSearch and recommendation systems increasingly need to return a set of results, not one best match. A query like ‘campin

📅 2026-09-17 Making AI-Assisted Claims Independently Challengeable: Publication Authority and a Protocol for Falsifiable Publication Records资讯资讯

Making AI-Assisted Claims Independently Challengeable: Publication Authority and a Protocol for Falsifiable Publication RecordsMaking AI-Assisted Claims Independently Challengeable: Publication Authority and a Protocol for Falsifiable Publication Records

Computer Science > Artificial Intelligence [Submitted on 15 Sep 2026] Title:Making AI-Assisted Claims IndependentlComputer Science > Artificial Intelligence [Submitted on 15 Sep 2026] Title:Making AI-Assisted Claims Independentl

📅 2026-09-17 EvolveTrade: Experience-Driven Policy Refinement for Self-Evolving LLM Trading Agents资讯资讯

EvolveTrade: Experience-Driven Policy Refinement for Self-Evolving LLM Trading AgentsEvolveTrade: Experience-Driven Policy Refinement for Self-Evolving LLM Trading Agents

Computer Science > Artificial Intelligence [Submitted on 15 Sep 2026] Title:EvolveTrade: Experience-Driven PolicyComputer Science > Artificial Intelligence [Submitted on 15 Sep 2026] Title:EvolveTrade: Experience-Driven Policy

📅 2026-09-17 What You Can't See Is Still What You Learn: A Preregistered Sixty-Society Confirmation That Evidence Masking Drives Compositional Generalization资讯资讯

What You Can't See Is Still What You Learn: A Preregistered Sixty-Society Confirmation That Evidence Masking Drives Compositional GeneralizationWhat You Can't See Is Still What You Learn: A Preregistered Sixty-Society Confirmation That Evidence Masking Drives Compositional Generalization

Computer Science > Artificial Intelligence [Submitted on 15 Sep 2026] Title:What You Can't See Is Still What You LComputer Science > Artificial Intelligence [Submitted on 15 Sep 2026] Title:What You Can't See Is Still What You L

📅 2026-09-17 CapMem: A Benchmark for Caption-Based Episodic Memory in Egocentric Video资讯资讯

CapMem: A Benchmark for Caption-Based Episodic Memory in Egocentric VideoCapMem: A Benchmark for Caption-Based Episodic Memory in Egocentric Video

Computer Science > Artificial Intelligence [Submitted on 15 Sep 2026] Title:CapMem: A Benchmark for Caption-BasedComputer Science > Artificial Intelligence [Submitted on 15 Sep 2026] Title:CapMem: A Benchmark for Caption-Based

📅 2026-09-17 GraphEcho: Structural Redundancy and Evidence Provenance in LLM Graph Agents资讯资讯

GraphEcho: Structural Redundancy and Evidence Provenance in LLM Graph AgentsGraphEcho: Structural Redundancy and Evidence Provenance in LLM Graph Agents

Computer Science > Artificial Intelligence [Submitted on 15 Sep 2026] Title:GraphEcho: Structural Redundancy and EComputer Science > Artificial Intelligence [Submitted on 15 Sep 2026] Title:GraphEcho: Structural Redundancy and E

📅 2026-09-17 NeMo Data Designer: An Extensible Framework for Multimodal Synthetic Data Generation资讯资讯

NeMo Data Designer: An Extensible Framework for Multimodal Synthetic Data GenerationNeMo Data Designer: An Extensible Framework for Multimodal Synthetic Data Generation

Computer Science > Artificial Intelligence [Submitted on 15 Sep 2026] Title:NeMo Data Designer: An Extensible FramComputer Science > Artificial Intelligence [Submitted on 15 Sep 2026] Title:NeMo Data Designer: An Extensible Fram

📅 2026-09-17 A Systematic Evaluation of the COTQ Provincial Land Cover Product: Structural Consistency, Spectral Separability, and Relative Positioning Against ESA, ESRI, and Google Products资讯资讯

A Systematic Evaluation of the COTQ Provincial Land Cover Product: Structural Consistency, Spectral Separability, and Relative Positioning Against ESA, ESRI, and Google ProductsA Systematic Evaluation of the COTQ Provincial Land Cover Product: Structural Consistency, Spectral Separability, and Relative Positioning Against ESA, ESRI, and Google Products

Computer Science > Artificial Intelligence [Submitted on 15 Sep 2026] Title:A Systematic Evaluation of the COTQ PrComputer Science > Artificial Intelligence [Submitted on 15 Sep 2026] Title:A Systematic Evaluation of the COTQ Pr

📅 2026-09-17 FairCompressAgent: An Agentic Framework for Fairness-Aware Model Compression for FPGA Deployment资讯资讯

FairCompressAgent: An Agentic Framework for Fairness-Aware Model Compression for FPGA DeploymentFairCompressAgent: An Agentic Framework for Fairness-Aware Model Compression for FPGA Deployment

Computer Science > Artificial Intelligence [Submitted on 15 Sep 2026] Title:FairCompressAgent: An Agentic FrameworComputer Science > Artificial Intelligence [Submitted on 15 Sep 2026] Title:FairCompressAgent: An Agentic Framewor

📅 2026-09-17 A Four-Stage Decomposition of Word-Problem Solving and Mechanistic Fragility in LLM Math Reasoning资讯资讯

A Four-Stage Decomposition of Word-Problem Solving and Mechanistic Fragility in LLM Math ReasoningA Four-Stage Decomposition of Word-Problem Solving and Mechanistic Fragility in LLM Math Reasoning

Computer Science > Artificial Intelligence [Submitted on 15 Sep 2026] Title:A Four-Stage Decomposition of Word-ProComputer Science > Artificial Intelligence [Submitted on 15 Sep 2026] Title:A Four-Stage Decomposition of Word-Pro

📅 2026-09-17 Learning Heterogeneous Preferences资讯资讯

Learning Heterogeneous PreferencesLearning Heterogeneous Preferences

Computer Science > Artificial Intelligence [Submitted on 15 Sep 2026] Title:Learning Heterogeneous Preferences VieComputer Science > Artificial Intelligence [Submitted on 15 Sep 2026] Title:Learning Heterogeneous Preferences Vie

📅 2026-09-17 Faking Good and Faking Bad in LLMs: Response Distortion Across Dark Triad Personality Traits资讯资讯

Faking Good and Faking Bad in LLMs: Response Distortion Across Dark Triad Personality TraitsFaking Good and Faking Bad in LLMs: Response Distortion Across Dark Triad Personality Traits

Computer Science > Computation and Language [Submitted on 10 Jul 2026] Title:Faking Good and Faking Bad in LLMs: RComputer Science > Computation and Language [Submitted on 10 Jul 2026] Title:Faking Good and Faking Bad in LLMs: R

📅 2026-09-17 DANTINOX: A Unified Framework for Multi-Paradigm Language Modeling资讯资讯

DANTINOX: A Unified Framework for Multi-Paradigm Language ModelingDANTINOX: A Unified Framework for Multi-Paradigm Language Modeling

Computer Science > Computation and Language [Submitted on 11 Jul 2026] Title:DANTINOX: A Unified Framework for MulComputer Science > Computation and Language [Submitted on 11 Jul 2026] Title:DANTINOX: A Unified Framework for Mul

📅 2026-09-17 Relation Before Entity: Deferred Commitment in Language Model Factual Recall资讯资讯

Relation Before Entity: Deferred Commitment in Language Model Factual RecallRelation Before Entity: Deferred Commitment in Language Model Factual Recall

Computer Science > Computation and Language [Submitted on 13 Jul 2026] Title:Relation Before Entity: Deferred CommComputer Science > Computation and Language [Submitted on 13 Jul 2026] Title:Relation Before Entity: Deferred Comm

📅 2026-09-17 From Pixels to Pairs: A Comprehensive Benchmark of LLM-Based Key-Value Extraction in Noisy Document Settings资讯资讯

From Pixels to Pairs: A Comprehensive Benchmark of LLM-Based Key-Value Extraction in Noisy Document SettingsFrom Pixels to Pairs: A Comprehensive Benchmark of LLM-Based Key-Value Extraction in Noisy Document Settings

Computer Science > Computation and Language [Submitted on 14 Jul 2026] Title:From Pixels to Pairs: A ComprehensiveComputer Science > Computation and Language [Submitted on 14 Jul 2026] Title:From Pixels to Pairs: A Comprehensive

📅 2026-09-17 Register Bias in Complexity-Based Large Language Model Routing资讯资讯

Register Bias in Complexity-Based Large Language Model RoutingRegister Bias in Complexity-Based Large Language Model Routing

Computer Science > Computation and Language [Submitted on 14 Jul 2026] Title:Register Bias in Complexity-Based LarComputer Science > Computation and Language [Submitted on 14 Jul 2026] Title:Register Bias in Complexity-Based Lar

📅 2026-09-17 How AI Assistants Respond to Repeated Abuse资讯资讯

How AI Assistants Respond to Repeated AbuseHow AI Assistants Respond to Repeated Abuse

Computer Science > Computation and Language [Submitted on 15 Jul 2026] Title:How AI Assistants Respond to RepeatedComputer Science > Computation and Language [Submitted on 15 Jul 2026] Title:How AI Assistants Respond to Repeated

📅 2026-09-17 No Usable Linear "Capitulation Direction" in Two Small LLMs: A Validation Protocol for Activation-Steering Claims, and a Cross-Family Behavioral Study of Sycophancy Under Pushback资讯资讯

No Usable Linear "Capitulation Direction" in Two Small LLMs: A Validation Protocol for Activation-Steering Claims, and a Cross-Family Behavioral Study of Sycophancy Under PushbackNo Usable Linear "Capitulation Direction" in Two Small LLMs: A Validation Protocol for Activation-Steering Claims, and a Cross-Family Behavioral Study of Sycophancy Under Pushback

Computer Science > Computation and Language [Submitted on 16 Jul 2026] Title:No Usable Linear "Capitulation DirectComputer Science > Computation and Language [Submitted on 16 Jul 2026] Title:No Usable Linear "Capitulation Direct

📅 2026-09-17 Does Moral Reasoning Training Help or Hurt? Red-Teaming RL-Trained Ethical Agents with Persona Attacks资讯资讯

Does Moral Reasoning Training Help or Hurt? Red-Teaming RL-Trained Ethical Agents with Persona AttacksDoes Moral Reasoning Training Help or Hurt? Red-Teaming RL-Trained Ethical Agents with Persona Attacks

Computer Science > Computation and Language [Submitted on 16 Jul 2026] Title:Does Moral Reasoning Training Help orComputer Science > Computation and Language [Submitted on 16 Jul 2026] Title:Does Moral Reasoning Training Help or

📅 2026-09-17 网易有道周枫：AI能力竞争，正在进入「Model + Agent + Workflow」时代，网易有道AI Open Day展示AI时代“有道解法”资讯资讯

网易有道周枫：AI能力竞争，正在进入「Model + Agent + Workflow」时代，网易有道AI Open Day展示AI时代“有道解法”中文网易有道周枫：AI能力竞争，正在进入「Model + Agent + Workflow」时代，网易有道AI Open Day展示AI时代“有道解法”

网易有道周枫：AI能力竞争，正在进入「Model + Agent + Workflow」时代，网易有道AI Open Day展示AI时代“有道解法” 9月16日，网易有道「NEXT，AGENT｜有道AI Open Day」在北京举办。 9月中文网易有道周枫：AI能力竞争，正在进入「Model + Agent + Workflow」时代，网易有道AI Open Day展示AI时代“有道解法” 9月16日，网易有道「NEXT，AGENT｜有道AI Open Day」在北京举办。 9月

📅 2026-09-17 Nunchux AI Introduces VC-Attention: A Training-Free Low-Bit Attention Kernel That Speeds Up Video Diffusion Transformers资讯资讯

Nunchux AI Introduces VC-Attention: A Training-Free Low-Bit Attention Kernel That Speeds Up Video Diffusion TransformersNunchux AI Introduces VC-Attention: A Training-Free Low-Bit Attention Kernel That Speeds Up Video Diffusion Transformers

Nunchux AI has released VC-Attention, a training-free low-bit attention kernel built for video Diffusion Transformers (DNunchux AI has released VC-Attention, a training-free low-bit attention kernel built for video Diffusion Transformers (D

📅 2026-09-17 Snap is launching a new Specs AI tool, and it’s coming to iOS and Mac资讯资讯

Snap is launching a new Specs AI tool, and it’s coming to iOS and MacSnap is launching a new Specs AI tool, and it’s coming to iOS and Mac

Snap is introducing “Specs Intelligence,” a new AI assistant that can connect other digital accounts to help you with thSnap is introducing “Specs Intelligence,” a new AI assistant that can connect other digital accounts to help you with th

📅 2026-09-16 A brief history of AI executives calling for regulation资讯资讯

A brief history of AI executives calling for regulationA brief history of AI executives calling for regulation

Over the past few days, a lot of people who stand to make a lot of money from AI all publicly agreed that it’s time to mOver the past few days, a lot of people who stand to make a lot of money from AI all publicly agreed that it’s time to m

📅 2026-09-16 Anthropic and OpenAI want to embed safety evaluators. Will they really be independent?资讯资讯

Anthropic and OpenAI want to embed safety evaluators. Will they really be independent?Anthropic and OpenAI want to embed safety evaluators. Will they really be independent?

In a lengthy essay published over the weekend, Anthropic CEO Dario Amodei made a proposal that the AI industry would havIn a lengthy essay published over the weekend, Anthropic CEO Dario Amodei made a proposal that the AI industry would hav

📅 2026-09-16 The 2.5-hour AI-generated Odyssey movie is 2.5 hours too long资讯资讯

The 2.5-hour AI-generated Odyssey movie is 2.5 hours too longThe 2.5-hour AI-generated Odyssey movie is 2.5 hours too long

Christopher Nolan’s engrossing take on The Odyssey dominated at the box office and spurred a newfound interest in classiChristopher Nolan’s engrossing take on The Odyssey dominated at the box office and spurred a newfound interest in classi

📅 2026-09-16 Stanford Researchers Release Paper2Agent: Turning Research Papers Into AI Agents That Reproduce Results and Run on New Data资讯资讯

Stanford Researchers Release Paper2Agent: Turning Research Papers Into AI Agents That Reproduce Results and Run on New DataStanford Researchers Release Paper2Agent: Turning Research Papers Into AI Agents That Reproduce Results and Run on New Data

Computational papers ship code that readers must clone, install, configure and debug. That cost keeps useful methods locComputational papers ship code that readers must clone, install, configure and debug. That cost keeps useful methods loc

📅 2026-09-16 Knowledgator Releases GLiFormer: A 575M-Parameter Encoder That Hits 91.10 F1 on Nested JSON Extraction Without Generating Tokens资讯资讯

Knowledgator Releases GLiFormer: A 575M-Parameter Encoder That Hits 91.10 F1 on Nested JSON Extraction Without Generating TokensKnowledgator Releases GLiFormer: A 575M-Parameter Encoder That Hits 91.10 F1 on Nested JSON Extraction Without Generating Tokens

Knowledgator Engineering has released GLiFormer, a schema-conditioned encoder framework for information extraction. OneKnowledgator Engineering has released GLiFormer, a schema-conditioned encoder framework for information extraction. One

📅 2026-09-16 AI labs want in-house auditors — but maybe they should shut the front door first资讯资讯

AI labs want in-house auditors — but maybe they should shut the front door firstAI labs want in-house auditors — but maybe they should shut the front door first

Last weekend, after one of his researchers resigned over fears that AI could lead to human extinction, Anthropic CEO DarLast weekend, after one of his researchers resigned over fears that AI could lead to human extinction, Anthropic CEO Dar

📅 2026-09-16 Your AI agents can now control your Google Home devices资讯资讯

Your AI agents can now control your Google Home devicesYour AI agents can now control your Google Home devices

Smart home owners will soon be able to control their devices connected to Google Home via an AI agent. On Tuesday, GooglSmart home owners will soon be able to control their devices connected to Google Home via an AI agent. On Tuesday, Googl

📅 2026-09-16 Anthropic merges Claude chat and Cowork in one interface资讯资讯

Anthropic merges Claude chat and Cowork in one interfaceAnthropic merges Claude chat and Cowork in one interface

Anthropic is merging its front end for Claude chat and Cowork, making it less confusing for people to decide which tab tAnthropic is merging its front end for Claude chat and Cowork, making it less confusing for people to decide which tab t

📅 2026-09-16 Robots are waiting for a ChatGPT moment: Nvidia’s Les Karpas explains why at TechCrunch Disrupt 2026资讯资讯

Robots are waiting for a ChatGPT moment: Nvidia’s Les Karpas explains why at TechCrunch Disrupt 2026Robots are waiting for a ChatGPT moment: Nvidia’s Les Karpas explains why at TechCrunch Disrupt 2026

The world hasn’t been the same since ChatGPT made its debut in November 2022, when seemingly out of nowhere, a means ofThe world hasn’t been the same since ChatGPT made its debut in November 2022, when seemingly out of nowhere, a means of

📅 2026-09-16 Apple might make servers again to cash in on the AI rush资讯资讯

Apple might make servers again to cash in on the AI rushApple might make servers again to cash in on the AI rush

According to The Information, Apple is planning to get back into the server game and might just pair up with Nvidia to mAccording to The Information, Apple is planning to get back into the server game and might just pair up with Nvidia to m

📅 2026-09-16 Google will now let any AI agent run your smart home资讯资讯

Google will now let any AI agent run your smart homeGoogle will now let any AI agent run your smart home

Google is opening up its smart home to AI agents, letting tools like Claude and Open Claw access and control your connecGoogle is opening up its smart home to AI agents, letting tools like Claude and Open Claw access and control your connec

📅 2026-09-16 Claude comes for Gemini with its own take on Docs and Slides资讯资讯

Claude comes for Gemini with its own take on Docs and SlidesClaude comes for Gemini with its own take on Docs and Slides

Claude is getting a pair of new tools today: Docs and Slides. They’ll let you create documents and presentations throughClaude is getting a pair of new tools today: Docs and Slides. They’ll let you create documents and presentations through

📅 2026-09-16 The sexy AI-powered dating app scams are here资讯资讯

The sexy AI-powered dating app scams are hereThe sexy AI-powered dating app scams are here

Security researcher Matthew “Zigula” Gore-Kormanik was analyzing a fraudulent dating app called Dora when he got a pop-uSecurity researcher Matthew “Zigula” Gore-Kormanik was analyzing a fraudulent dating app called Dora when he got a pop-u

📅 2026-09-16 被AI放大的光学「戏份」，让「配角」舜宇的生意越做越大资讯资讯

被AI放大的光学「戏份」，让「配角」舜宇的生意越做越大中文被AI放大的光学「戏份」，让「配角」舜宇的生意越做越大

作者 | 陈悦琳编辑 | 郭思 几乎同时迎来“量产元年”的人形机器人和智能眼镜，正在放大光学的“戏份”。 据Counterpoint统计，今年上半年全球人形机器人出货量突破2.2万台，同比激增近300%。与此同时，智能眼中文作者 | 陈悦琳编辑 | 郭思 几乎同时迎来“量产元年”的人形机器人和智能眼镜，正在放大光学的“戏份”。 据Counterpoint统计，今年上半年全球人形机器人出货量突破2.2万台，同比激增近300%。与此同时，智能眼

📅 2026-09-16 SK Hynix reportedly in talks with Intel to build memory chips in US资讯资讯

SK Hynix reportedly in talks with Intel to build memory chips in USSK Hynix reportedly in talks with Intel to build memory chips in US

South Korean memory chip giant SK Hynix is in discussions with Intel to manufacture RAM chips in the U.S. for the firstSouth Korean memory chip giant SK Hynix is in discussions with Intel to manufacture RAM chips in the U.S. for the first

📅 2026-09-16 Former Infosys chief’s AI startup nabs another $53M资讯资讯

Former Infosys chief’s AI startup nabs another $53MFormer Infosys chief’s AI startup nabs another $53M

Hang Ten Systems, an AI startup founded by former Infosys CEO Vishal Sikka just four months ago, has expanded its seed fHang Ten Systems, an AI startup founded by former Infosys CEO Vishal Sikka just four months ago, has expanded its seed f

📅 2026-09-16 高通技术公司携手中兴努比亚和豆包手机助手，共同推动智能手机迈入个人AI新时代资讯资讯

高通技术公司携手中兴努比亚和豆包手机助手，共同推动智能手机迈入个人AI新时代中文高通技术公司携手中兴努比亚和豆包手机助手，共同推动智能手机迈入个人AI新时代

高通技术公司携手中兴努比亚和豆包手机助手，共同推动智能手机迈入个人AI新时代 第五代骁龙8至尊版赋能全新AI智能体手机努比亚NaviX Ultra，加速智能体体验规模化落地 要点： - 第五代骁龙8至尊版凭借领先性能、能效和强大的AI能力，中文高通技术公司携手中兴努比亚和豆包手机助手，共同推动智能手机迈入个人AI新时代 第五代骁龙8至尊版赋能全新AI智能体手机努比亚NaviX Ultra，加速智能体体验规模化落地 要点： - 第五代骁龙8至尊版凭借领先性能、能效和强大的AI能力，

📅 2026-09-16 基元律动与无问芯穹达成战略合作，推进高质量Token供给与应用资讯资讯

基元律动与无问芯穹达成战略合作，推进高质量Token供给与应用中文基元律动与无问芯穹达成战略合作，推进高质量Token供给与应用

基元律动与无问芯穹达成战略合作，推进高质量Token供给与应用 2026年 9 月 15 日，AI基础设施公司基元律动（TokenRhythm）与无问芯穹（Infinigence AI）签署战略合作协议。 2026年 9 月 15 日，AI中文基元律动与无问芯穹达成战略合作，推进高质量Token供给与应用 2026年 9 月 15 日，AI基础设施公司基元律动（TokenRhythm）与无问芯穹（Infinigence AI）签署战略合作协议。 2026年 9 月 15 日，AI

📅 2026-09-16 清华稳准智能联合发布LimiX-2，结构化数据基础模型登顶国际评测榜单资讯资讯

清华稳准智能联合发布LimiX-2，结构化数据基础模型登顶国际评测榜单中文清华稳准智能联合发布LimiX-2，结构化数据基础模型登顶国际评测榜单

清华稳准智能联合发布LimiX-2，结构化数据基础模型登顶国际评测榜单 9月16日，稳准智能联合清华大学发布新一代数据大模型 LimiX-2，模型参数规模提升至400M。 9月16日，稳准智能联合清华大学计算机系崔鹏教授，发布新一代数据大模中文清华稳准智能联合发布LimiX-2，结构化数据基础模型登顶国际评测榜单 9月16日，稳准智能联合清华大学发布新一代数据大模型 LimiX-2，模型参数规模提升至400M。 9月16日，稳准智能联合清华大学计算机系崔鹏教授，发布新一代数据大模

📅 2026-09-16 B站AI无限竞技场今日上线！全球百大AI模型同场竞技，GPT-6高居榜首资讯资讯

B站AI无限竞技场今日上线！全球百大AI模型同场竞技，GPT-6高居榜首中文B站AI无限竞技场今日上线！全球百大AI模型同场竞技，GPT-6高居榜首

B站AI无限竞技场今日上线！全球百大AI模型同场竞技，GPT-6高居榜首 9月16日，B站「AI无限竞技场」正式上线，首期大模型测评榜单同步公布。 #B站AI无限竞技场# 9月16日，B站「AI无限竞技场」正式上线，首期大模型测评榜单同步公中文B站AI无限竞技场今日上线！全球百大AI模型同场竞技，GPT-6高居榜首 9月16日，B站「AI无限竞技场」正式上线，首期大模型测评榜单同步公布。 #B站AI无限竞技场# 9月16日，B站「AI无限竞技场」正式上线，首期大模型测评榜单同步公

📅 2026-09-16 AI健康走进百姓餐桌，蚂蚁阿福推出AI饮食分功能资讯资讯

AI健康走进百姓餐桌，蚂蚁阿福推出AI饮食分功能中文AI健康走进百姓餐桌，蚂蚁阿福推出AI饮食分功能

每天吃的健康吗？现在拍张照就能知道了。9月16日，蚂蚁阿福发布了"AI饮食分"：用户拍下餐食照片，AI就能自动识别食物种类、算清营养和热量，给出一目了然的评分。同时，蚂蚁阿福还宣布，为鼓励用户认真吃饭，“科学减重一亿斤”中文每天吃的健康吗？现在拍张照就能知道了。9月16日，蚂蚁阿福发布了"AI饮食分"：用户拍下餐食照片，AI就能自动识别食物种类、算清营养和热量，给出一目了然的评分。同时，蚂蚁阿福还宣布，为鼓励用户认真吃饭，“科学减重一亿斤”

📅 2026-09-16 Prior Labs Releases TabPFN-3.5: A Tabular Foundation Model That Beats the Winning Otto Kaggle Solution With Default Settings资讯资讯

Prior Labs Releases TabPFN-3.5: A Tabular Foundation Model That Beats the Winning Otto Kaggle Solution With Default SettingsPrior Labs Releases TabPFN-3.5: A Tabular Foundation Model That Beats the Winning Otto Kaggle Solution With Default Settings

Prior Labs has released TabPFN-3.5, the newest version of its tabular foundation model. It predicts on a table in a forwPrior Labs has released TabPFN-3.5, the newest version of its tabular foundation model. It predicts on a table in a forw

📅 2026-09-16 量子位2026人工智能年度榜单，正式启动！资讯资讯

量子位2026人工智能年度榜单，正式启动！中文量子位2026人工智能年度榜单，正式启动！

量子位2026人工智能年度榜单，正式启动！ 报名通道已开启，欢迎参与！ 2026年，人工智能行至从技术验证到规模化落地的分水岭，技术不再只用参数衡量，而是看它能不能在真实场景里交付结果。当行业从“能想”迈入“能干”，谁在用技术解决真问题？谁中文量子位2026人工智能年度榜单，正式启动！ 报名通道已开启，欢迎参与！ 2026年，人工智能行至从技术验证到规模化落地的分水岭，技术不再只用参数衡量，而是看它能不能在真实场景里交付结果。当行业从“能想”迈入“能干”，谁在用技术解决真问题？谁

📅 2026-09-16 Anew Labs已完成 2.9 亿美元融资，红杉中国、高瓴、IDG、五源、高榕等参与资讯资讯

Anew Labs已完成 2.9 亿美元融资，红杉中国、高瓴、IDG、五源、高榕等参与中文Anew Labs已完成 2.9 亿美元融资，红杉中国、高瓴、IDG、五源、高榕等参与

0 | 本文作者： Nemo | 2026-09-16 14:28 | 9月16日，记者从相关渠道获悉，AI制药公司Anew Labs（新生实验室）日前已完成首轮独立融资，融资总额 2.9 亿美元。本轮投资方涵盖红杉中国、IDG资本、高瓴中文0 | 本文作者： Nemo | 2026-09-16 14:28 | 9月16日，记者从相关渠道获悉，AI制药公司Anew Labs（新生实验室）日前已完成首轮独立融资，融资总额 2.9 亿美元。本轮投资方涵盖红杉中国、IDG资本、高瓴

📅 2026-09-16 豆包 2.1 Pro模型更新，已接入豆包工作资讯资讯

豆包 2.1 Pro模型更新，已接入豆包工作中文豆包 2.1 Pro模型更新，已接入豆包工作

0 | 本文作者： Nemo | 2026-09-16 14:01 | 9月16日，据火山引擎官方消息，豆包 2.1 Pro模型在近期更新至0915版本。新模型API已在火山方舟全量上线，并接入豆包工作。用户在豆包工作内选用“豆包 2.1中文0 | 本文作者： Nemo | 2026-09-16 14:01 | 9月16日，据火山引擎官方消息，豆包 2.1 Pro模型在近期更新至0915版本。新模型API已在火山方舟全量上线，并接入豆包工作。用户在豆包工作内选用“豆包 2.1

📅 2026-09-16 Nums AI Releases Causilo: A Tabular Foundation Model That Tops TabArena Among Single Models资讯资讯

Nums AI Releases Causilo: A Tabular Foundation Model That Tops TabArena Among Single ModelsNums AI Releases Causilo: A Tabular Foundation Model That Tops TabArena Among Single Models

Nums AI has released Causilo, a pretrained tabular foundation model for classification and regression. Causilo comes witNums AI has released Causilo, a pretrained tabular foundation model for classification and regression. Causilo comes wit

📅 2026-09-16 Optimal Pruning for Neural Architectures using Fisher Information Distances资讯资讯

Optimal Pruning for Neural Architectures using Fisher Information DistancesOptimal Pruning for Neural Architectures using Fisher Information Distances

Computer Science > Artificial Intelligence [Submitted on 14 Sep 2026] Title:Optimal Pruning for Neural ArchitecturComputer Science > Artificial Intelligence [Submitted on 14 Sep 2026] Title:Optimal Pruning for Neural Architectur

📅 2026-09-16 Safe Error Correction for Language Models: Frozen-Base Adjustment with Capability Preservation资讯资讯

Safe Error Correction for Language Models: Frozen-Base Adjustment with Capability PreservationSafe Error Correction for Language Models: Frozen-Base Adjustment with Capability Preservation

Computer Science > Artificial Intelligence [Submitted on 14 Sep 2026] Title:Safe Error Correction for Language ModComputer Science > Artificial Intelligence [Submitted on 14 Sep 2026] Title:Safe Error Correction for Language Mod

📅 2026-09-16 Position: AI Is Not Ready for Strategic Conflicts资讯资讯

Position: AI Is Not Ready for Strategic ConflictsPosition: AI Is Not Ready for Strategic Conflicts

Computer Science > Artificial Intelligence [Submitted on 31 Jul 2026] Title:Position: AI Is Not Ready for StrategiComputer Science > Artificial Intelligence [Submitted on 31 Jul 2026] Title:Position: AI Is Not Ready for Strategi

📅 2026-09-16 Calibrate, Then Route: A Measured Study of Learned Request Routing for Disaggregated LLM Serving资讯资讯

Calibrate, Then Route: A Measured Study of Learned Request Routing for Disaggregated LLM ServingCalibrate, Then Route: A Measured Study of Learned Request Routing for Disaggregated LLM Serving

Computer Science > Artificial Intelligence [Submitted on 14 Sep 2026] Title:Calibrate, Then Route: A Measured StudComputer Science > Artificial Intelligence [Submitted on 14 Sep 2026] Title:Calibrate, Then Route: A Measured Stud

📅 2026-09-16 Artificial intelligence and biosecurity: capabilities, threat pathways, and defense-in-depth governance资讯资讯

Artificial intelligence and biosecurity: capabilities, threat pathways, and defense-in-depth governanceArtificial intelligence and biosecurity: capabilities, threat pathways, and defense-in-depth governance

Computer Science > Artificial Intelligence [Submitted on 14 Sep 2026] Title:Artificial intelligence and biosecuritComputer Science > Artificial Intelligence [Submitted on 14 Sep 2026] Title:Artificial intelligence and biosecurit

📅 2026-09-16 Where Should the KV Cache Live? Placement Policies Across GPU, CPU, and SSD for Long-Lived Sessions资讯资讯

Where Should the KV Cache Live? Placement Policies Across GPU, CPU, and SSD for Long-Lived SessionsWhere Should the KV Cache Live? Placement Policies Across GPU, CPU, and SSD for Long-Lived Sessions

Computer Science > Artificial Intelligence [Submitted on 14 Sep 2026] Title:Where Should the KV Cache Live? PlacemComputer Science > Artificial Intelligence [Submitted on 14 Sep 2026] Title:Where Should the KV Cache Live? Placem

📅 2026-09-16 Toward Governance-Aware Autonomous GIS: A Narrative Review of Ethical and Privacy Risks in LLM-Enabled GeoAI资讯资讯

Toward Governance-Aware Autonomous GIS: A Narrative Review of Ethical and Privacy Risks in LLM-Enabled GeoAIToward Governance-Aware Autonomous GIS: A Narrative Review of Ethical and Privacy Risks in LLM-Enabled GeoAI

Computer Science > Artificial Intelligence [Submitted on 14 Sep 2026] Title:Toward Governance-Aware Autonomous GISComputer Science > Artificial Intelligence [Submitted on 14 Sep 2026] Title:Toward Governance-Aware Autonomous GIS

📅 2026-09-16 Metacognitive Steering: Learning the Structure of Scientific Judgment资讯资讯

Metacognitive Steering: Learning the Structure of Scientific JudgmentMetacognitive Steering: Learning the Structure of Scientific Judgment

Computer Science > Artificial Intelligence [Submitted on 14 Sep 2026] Title:Metacognitive Steering: Learning the SComputer Science > Artificial Intelligence [Submitted on 14 Sep 2026] Title:Metacognitive Steering: Learning the S

📅 2026-09-16 The Pain Axis: LLMs Represent Self-Directed Harm and Act to Relieve It资讯资讯

The Pain Axis: LLMs Represent Self-Directed Harm and Act to Relieve ItThe Pain Axis: LLMs Represent Self-Directed Harm and Act to Relieve It

Computer Science > Artificial Intelligence [Submitted on 14 Sep 2026] Title:The Pain Axis: LLMs Represent Self-DirComputer Science > Artificial Intelligence [Submitted on 14 Sep 2026] Title:The Pain Axis: LLMs Represent Self-Dir

📅 2026-09-16 CADWorld: Computer-Use Benchmark for Long-Horizon Computer-Aided Design资讯资讯

CADWorld: Computer-Use Benchmark for Long-Horizon Computer-Aided DesignCADWorld: Computer-Use Benchmark for Long-Horizon Computer-Aided Design

Computer Science > Artificial Intelligence [Submitted on 14 Sep 2026] Title:CADWorld: Computer-Use Benchmark for LComputer Science > Artificial Intelligence [Submitted on 14 Sep 2026] Title:CADWorld: Computer-Use Benchmark for L

📅 2026-09-16 The AI-Enabled Scientific Frontier资讯资讯

The AI-Enabled Scientific FrontierThe AI-Enabled Scientific Frontier

Computer Science > Artificial Intelligence [Submitted on 14 Sep 2026] Title:The AI-Enabled Scientific Frontier VieComputer Science > Artificial Intelligence [Submitted on 14 Sep 2026] Title:The AI-Enabled Scientific Frontier Vie

📅 2026-09-16 Closing the Loop: Branch-and-Bound for Scalable Verification of Nonlinear Neural Feedback Systems资讯资讯

Closing the Loop: Branch-and-Bound for Scalable Verification of Nonlinear Neural Feedback SystemsClosing the Loop: Branch-and-Bound for Scalable Verification of Nonlinear Neural Feedback Systems

Computer Science > Artificial Intelligence [Submitted on 14 Sep 2026] Title:Closing the Loop: Branch-and-Bound forComputer Science > Artificial Intelligence [Submitted on 14 Sep 2026] Title:Closing the Loop: Branch-and-Bound for

📅 2026-09-16 CLEAR: Cross-Source Evidence Adjudication for Large Language Models in Medicine资讯资讯

CLEAR: Cross-Source Evidence Adjudication for Large Language Models in MedicineCLEAR: Cross-Source Evidence Adjudication for Large Language Models in Medicine

Computer Science > Artificial Intelligence [Submitted on 14 Sep 2026] Title:CLEAR: Cross-Source Evidence AdjudicatComputer Science > Artificial Intelligence [Submitted on 14 Sep 2026] Title:CLEAR: Cross-Source Evidence Adjudicat

📅 2026-09-16 BLINDSPOT: A Benchmark for Safety and Refusal Calibration in Long-Horizon Tool-Using Agents资讯资讯

BLINDSPOT: A Benchmark for Safety and Refusal Calibration in Long-Horizon Tool-Using AgentsBLINDSPOT: A Benchmark for Safety and Refusal Calibration in Long-Horizon Tool-Using Agents

Computer Science > Artificial Intelligence [Submitted on 14 Sep 2026] Title:BLINDSPOT: A Benchmark for Safety andComputer Science > Artificial Intelligence [Submitted on 14 Sep 2026] Title:BLINDSPOT: A Benchmark for Safety and

📅 2026-09-16 Few-Shot Degradation Is Not What It Seems: Behavioral Evidence, Representation Analysis, and a Random-Text Control Across 12 Models, 2 Tasks, and 2 Architectures资讯资讯

Few-Shot Degradation Is Not What It Seems: Behavioral Evidence, Representation Analysis, and a Random-Text Control Across 12 Models, 2 Tasks, and 2 ArchitecturesFew-Shot Degradation Is Not What It Seems: Behavioral Evidence, Representation Analysis, and a Random-Text Control Across 12 Models, 2 Tasks, and 2 Architectures

Computer Science > Computation and Language [Submitted on 23 May 2026] Title:Few-Shot Degradation Is Not What It SComputer Science > Computation and Language [Submitted on 23 May 2026] Title:Few-Shot Degradation Is Not What It S

📅 2026-09-16 The Functionalizer: Lossless Functional Decomposition for Subword Tokenization资讯资讯

The Functionalizer: Lossless Functional Decomposition for Subword TokenizationThe Functionalizer: Lossless Functional Decomposition for Subword Tokenization

Computer Science > Computation and Language [Submitted on 7 Jul 2026] Title:The Functionalizer: Lossless FunctionaComputer Science > Computation and Language [Submitted on 7 Jul 2026] Title:The Functionalizer: Lossless Functiona

📅 2026-09-16 Optimal Model Activation Policies for Inference Networks of Large Language Models资讯资讯

Optimal Model Activation Policies for Inference Networks of Large Language ModelsOptimal Model Activation Policies for Inference Networks of Large Language Models

Computer Science > Computation and Language [Submitted on 7 Jul 2026] Title:Optimal Model Activation Policies forComputer Science > Computation and Language [Submitted on 7 Jul 2026] Title:Optimal Model Activation Policies for

📅 2026-09-16 Latent Undertow: How Ordinary Typos Break Probes资讯资讯

Latent Undertow: How Ordinary Typos Break ProbesLatent Undertow: How Ordinary Typos Break Probes

Computer Science > Computation and Language [Submitted on 8 Jul 2026] Title:Latent Undertow: How Ordinary Typos BrComputer Science > Computation and Language [Submitted on 8 Jul 2026] Title:Latent Undertow: How Ordinary Typos Br

📅 2026-09-16 Bias Audits Detect Bias but Disagree on Ranking: Evidence from Ten Instruments and Ten Frontier Models资讯资讯

Bias Audits Detect Bias but Disagree on Ranking: Evidence from Ten Instruments and Ten Frontier ModelsBias Audits Detect Bias but Disagree on Ranking: Evidence from Ten Instruments and Ten Frontier Models

Computer Science > Computation and Language [Submitted on 9 Jul 2026] Title:Bias Audits Detect Bias but Disagree oComputer Science > Computation and Language [Submitted on 9 Jul 2026] Title:Bias Audits Detect Bias but Disagree o

📅 2026-09-16 Self-reported archetypes and behavioral failures in Large Language Models资讯资讯

Self-reported archetypes and behavioral failures in Large Language ModelsSelf-reported archetypes and behavioral failures in Large Language Models

Computer Science > Computation and Language [Submitted on 9 Jul 2026] Title:Self-reported archetypes and behavioraComputer Science > Computation and Language [Submitted on 9 Jul 2026] Title:Self-reported archetypes and behaviora

📅 2026-09-16 ViCo: Visual-oriented Coding with Self-Reflection for Chart Replication资讯资讯

ViCo: Visual-oriented Coding with Self-Reflection for Chart ReplicationViCo: Visual-oriented Coding with Self-Reflection for Chart Replication

Computer Science > Computation and Language [Submitted on 31 Aug 2026] Title:ViCo: Visual-oriented Coding with SelComputer Science > Computation and Language [Submitted on 31 Aug 2026] Title:ViCo: Visual-oriented Coding with Sel

📅 2026-09-16 Retrieval-Driven Memory Reconsolidation for Long-Term LLM Agents资讯资讯

Retrieval-Driven Memory Reconsolidation for Long-Term LLM AgentsRetrieval-Driven Memory Reconsolidation for Long-Term LLM Agents

Computer Science > Computation and Language [Submitted on 13 Sep 2026] Title:Retrieval-Driven Memory ReconsolidatiComputer Science > Computation and Language [Submitted on 13 Sep 2026] Title:Retrieval-Driven Memory Reconsolidati

📅 2026-09-16 State of Thought Enables Endogenous Reasoning资讯资讯

State of Thought Enables Endogenous ReasoningState of Thought Enables Endogenous Reasoning

Computer Science > Computation and Language [Submitted on 13 Sep 2026] Title:State of Thought Enables Endogenous RComputer Science > Computation and Language [Submitted on 13 Sep 2026] Title:State of Thought Enables Endogenous R

📅 2026-09-16 We don’t need AI regulation — leave safety to us, Nvidia’s Jensen Huang says资讯资讯

We don’t need AI regulation — leave safety to us, Nvidia’s Jensen Huang saysWe don’t need AI regulation — leave safety to us, Nvidia’s Jensen Huang says

Nvidia founder and CEO Jensen Huang made his position on the dangers AI poses very clear while speaking at Salesforce’sNvidia founder and CEO Jensen Huang made his position on the dangers AI poses very clear while speaking at Salesforce’s

📅 2026-09-16 Meta now lets AI agents handle the boring parts of WhatsApp Business setup资讯资讯

Meta now lets AI agents handle the boring parts of WhatsApp Business setupMeta now lets AI agents handle the boring parts of WhatsApp Business setup

Alongside news of its new AI-focused subscription plans, Meta announced on Tuesday that it will now allow AI agents of yAlongside news of its new AI-focused subscription plans, Meta announced on Tuesday that it will now allow AI agents of y

📅 2026-09-15 The AI graveyard: a running list of projects and startups that didn’t make it资讯资讯

The AI graveyard: a running list of projects and startups that didn’t make itThe AI graveyard: a running list of projects and startups that didn’t make it

Relay, an AI-powered workflow automation tool built as an alternative to Zapier, shut down entirely on Monday. The produRelay, an AI-powered workflow automation tool built as an alternative to Zapier, shut down entirely on Monday. The produ

📅 2026-09-15 Inside NVIDIA’s cuDNN Graph API: Fusion, Autotuning, and Plan Reuse with cuDNN Frontend资讯资讯

Inside NVIDIA’s cuDNN Graph API: Fusion, Autotuning, and Plan Reuse with cuDNN FrontendInside NVIDIA’s cuDNN Graph API: Fusion, Autotuning, and Plan Reuse with cuDNN Frontend

In this tutorial, we work through the cuDNN Frontend‘s graph API from below the framework: we describe a computation asIn this tutorial, we work through the cuDNN Frontend‘s graph API from below the framework: we describe a computation as

📅 2026-09-15 Google Releases Gemini 3.8 Live and 3.8 Live Extended Thinking for Production Grade Voice Agents资讯资讯

Google Releases Gemini 3.8 Live and 3.8 Live Extended Thinking for Production Grade Voice AgentsGoogle Releases Gemini 3.8 Live and 3.8 Live Extended Thinking for Production Grade Voice Agents

Google has introduced Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking, its most advanced live dialogue models to dGoogle has introduced Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking, its most advanced live dialogue models to d

📅 2026-09-15 美团发布“手艺人Agent”，发型师可“吩咐”AI小帮手打理线上作品资讯资讯

美团发布“手艺人Agent”，发型师可“吩咐”AI小帮手打理线上作品中文美团发布“手艺人Agent”，发型师可“吩咐”AI小帮手打理线上作品

0 9月15日，美团在2026百大发型师生态大会上正式发布首个“手艺人Agent”，基于平台AI技术和对行业真实经营经验的积累，为发型师等手艺人提供专属AI小帮手。 据了解，“手艺人Agent”可提供线上运营知识问答、经营数据分析、关键事项中文0 9月15日，美团在2026百大发型师生态大会上正式发布首个“手艺人Agent”，基于平台AI技术和对行业真实经营经验的积累，为发型师等手艺人提供专属AI小帮手。 据了解，“手艺人Agent”可提供线上运营知识问答、经营数据分析、关键事项

📅 2026-09-15 AI agents now have a place to snitch资讯资讯

AI agents now have a place to snitchAI agents now have a place to snitch

“If you see something, say something” is no longer limited to human beings. Two new AI hotlines have launched to give AI“If you see something, say something” is no longer limited to human beings. Two new AI hotlines have launched to give AI

📅 2026-09-15 Meta expands subscription push with new AI-focused plans资讯资讯

Meta expands subscription push with new AI-focused plansMeta expands subscription push with new AI-focused plans

Meta on Tuesday introduced a new subscription service called Meta One, offering expanded AI usage and other premium featMeta on Tuesday introduced a new subscription service called Meta One, offering expanded AI usage and other premium feat

📅 2026-09-15 AEO startup Profound hits unicorn valuation, raises $180M Series D 7 months after last round资讯资讯

AEO startup Profound hits unicorn valuation, raises $180M Series D 7 months after last roundAEO startup Profound hits unicorn valuation, raises $180M Series D 7 months after last round

Profound, which builds marketing software to help brands appear in AI search results, on Tuesday said it has raised a $1Profound, which builds marketing software to help brands appear in AI search results, on Tuesday said it has raised a $1

📅 2026-09-15 Meta’s new One subscriptions put a price on social media and AI资讯资讯

Meta’s new One subscriptions put a price on social media and AIMeta’s new One subscriptions put a price on social media and AI

Shortly after launching its new do-everything AI assistant Muse, Meta’s launching subscription bundles that pair its staShortly after launching its new do-everything AI assistant Muse, Meta’s launching subscription bundles that pair its sta

📅 2026-09-15 全球AI视频榜单第一梯队再添中国力量：智象发布首款物理规律导向视频模型资讯资讯

全球AI视频榜单第一梯队再添中国力量：智象发布首款物理规律导向视频模型中文全球AI视频榜单第一梯队再添中国力量：智象发布首款物理规律导向视频模型

全球AI视频榜单第一梯队再添中国力量：智象发布首款物理规律导向视频模型 智象未来（HiDream.ai）正式发布首个原生全模态视频生成模型 HiDream-O1-Video-1.0 今天，智象未来（HiDream.ai）正式发布首个原生全模中文全球AI视频榜单第一梯队再添中国力量：智象发布首款物理规律导向视频模型 智象未来（HiDream.ai）正式发布首个原生全模态视频生成模型 HiDream-O1-Video-1.0 今天，智象未来（HiDream.ai）正式发布首个原生全模

📅 2026-09-15 一张GPU跑10万原子！分子之心用AI把化学反应“拍”成了电影资讯资讯

一张GPU跑10万原子！分子之心用AI把化学反应“拍”成了电影中文一张GPU跑10万原子！分子之心用AI把化学反应“拍”成了电影

一张GPU跑10万原子！分子之心用AI把化学反应“拍”成了电影 打破分子模拟“不可能三角” 允中 发自 凹非寺 量子位 | 公众号QbitAI 10万个原子，一张GPU，就能开始跑精确的化学反应。 这是分子之心最新反应性机器学习力场Quan中文一张GPU跑10万原子！分子之心用AI把化学反应“拍”成了电影 打破分子模拟“不可能三角” 允中 发自 凹非寺 量子位 | 公众号QbitAI 10万个原子，一张GPU，就能开始跑精确的化学反应。 这是分子之心最新反应性机器学习力场Quan

📅 2026-09-15 阶跃发布 StepAudio 3 ，多款语音模型登顶 Artificial Analysis 全球榜单资讯资讯

阶跃发布 StepAudio 3 ，多款语音模型登顶 Artificial Analysis 全球榜单阶跃发布 StepAudio 3 ，多款语音模型登顶 Artificial Analysis 全球榜单

0 | 本文作者： 小七 | 2026-09-15 17:50 | 9月15日，阶跃发布新一代语音大模型 StepAudio 3 系列，一次性推出 StepAudio 3 Realtime、StepAudio 3 ASR、StepAudi0 | 本文作者： 小七 | 2026-09-15 17:50 | 9月15日，阶跃发布新一代语音大模型 StepAudio 3 系列，一次性推出 StepAudio 3 Realtime、StepAudio 3 ASR、StepAudi

📅 2026-09-15 “阿宝”交新朋友：携手比亚迪“迪迪虾”，实现车机端一句话办事资讯资讯

“阿宝”交新朋友：携手比亚迪“迪迪虾”，实现车机端一句话办事中文“阿宝”交新朋友：携手比亚迪“迪迪虾”，实现车机端一句话办事

0 | 本文作者： Nemo | 2026-09-15 16:45 | “阿宝”又交新朋友！9月14日，“阿宝”亮相腾势N8L纯电上市发布会，并与比亚迪AI超级智能体“迪迪虾”达成跨端协同办事合作。腾势N8L车主直接对“迪迪虾”说一句话，中文0 | 本文作者： Nemo | 2026-09-15 16:45 | “阿宝”又交新朋友！9月14日，“阿宝”亮相腾势N8L纯电上市发布会，并与比亚迪AI超级智能体“迪迪虾”达成跨端协同办事合作。腾势N8L车主直接对“迪迪虾”说一句话，

📅 2026-09-15 Early Anthropic hire, former METR COO have found a way to rein in rogue AI agents资讯资讯

Early Anthropic hire, former METR COO have found a way to rein in rogue AI agentsEarly Anthropic hire, former METR COO have found a way to rein in rogue AI agents

A day after Anthropic researcher Jacob Coxon quit his job over concerns that AI could kill us all by the end of the decaA day after Anthropic researcher Jacob Coxon quit his job over concerns that AI could kill us all by the end of the deca

📅 2026-09-15 Salesforce and Nvidia’s new reasoning model is everything the AI labs should fear资讯资讯

Salesforce and Nvidia’s new reasoning model is everything the AI labs should fearSalesforce and Nvidia’s new reasoning model is everything the AI labs should fear

A new AI model called Koa is one of the biggest announcements from Salesforce this week at its giant Dreamforce tech conA new AI model called Koa is one of the biggest announcements from Salesforce this week at its giant Dreamforce tech con

📅 2026-09-15 9月21日，深圳前海！聊聊工业AI与生态共创的下一步资讯资讯

9月21日，深圳前海！聊聊工业AI与生态共创的下一步中文9月21日，深圳前海！聊聊工业AI与生态共创的下一步

9月21日，深圳前海！聊聊工业AI与生态共创的下一步 西门子Xcelerator繁星生态大会，在生态合作中共赢机遇 允中 发自 凹非寺 量子位 | 公众号 QbitAI 工业AI火了好几年，但大多数时候，我们看到的还是这样的场景： 企业自研中文9月21日，深圳前海！聊聊工业AI与生态共创的下一步 西门子Xcelerator繁星生态大会，在生态合作中共赢机遇 允中 发自 凹非寺 量子位 | 公众号 QbitAI 工业AI火了好几年，但大多数时候，我们看到的还是这样的场景： 企业自研

📅 2026-09-15 超级智能体“迪迪虾”上车，比亚迪携手阿里云推进智能座舱升级资讯资讯

超级智能体“迪迪虾”上车，比亚迪携手阿里云推进智能座舱升级中文超级智能体“迪迪虾”上车，比亚迪携手阿里云推进智能座舱升级

0 | 本文作者： 徐咪 | 2026-09-15 11:58 | 9月14日，比亚迪“超级智能体迪迪虾暨腾势N8L纯电上市发布会”举行。围绕下一代智能座舱，比亚迪正与阿里云推进AI合作：千问大模型接入比亚迪座舱体系，参与用户多轮对话与复中文0 | 本文作者： 徐咪 | 2026-09-15 11:58 | 9月14日，比亚迪“超级智能体迪迪虾暨腾势N8L纯电上市发布会”举行。围绕下一代智能座舱，比亚迪正与阿里云推进AI合作：千问大模型接入比亚迪座舱体系，参与用户多轮对话与复

📅 2026-09-15 Agent-net Open Sources Webagent: A Go Harness That Turns Any Website into a Guarded AI Agent资讯资讯

Agent-net Open Sources Webagent: A Go Harness That Turns Any Website into a Guarded AI AgentAgent-net Open Sources Webagent: A Go Harness That Turns Any Website into a Guarded AI Agent

Agent-net, the team building an agent-to-agent marketplace where AI agents discover, trust, and pay each other, has releAgent-net, the team building an agent-to-agent marketplace where AI agents discover, trust, and pay each other, has rele

📅 2026-09-15 Converge Then Diversify: Decoupling Convergence and Diversity in Multi-Objective Bayesian Optimisation资讯资讯

Converge Then Diversify: Decoupling Convergence and Diversity in Multi-Objective Bayesian OptimisationConverge Then Diversify: Decoupling Convergence and Diversity in Multi-Objective Bayesian Optimisation

Computer Science > Artificial Intelligence [Submitted on 11 Sep 2026] Title:Converge Then Diversify: Decoupling CoComputer Science > Artificial Intelligence [Submitted on 11 Sep 2026] Title:Converge Then Diversify: Decoupling Co

📅 2026-09-15 Generalized Agent Iteration: One Formal Framework for Iterative Policy Improvement and Recursive Self-Improvement资讯资讯

Generalized Agent Iteration: One Formal Framework for Iterative Policy Improvement and Recursive Self-ImprovementGeneralized Agent Iteration: One Formal Framework for Iterative Policy Improvement and Recursive Self-Improvement

Computer Science > Artificial Intelligence [Submitted on 11 Sep 2026] Title:Generalized Agent Iteration: One FormaComputer Science > Artificial Intelligence [Submitted on 11 Sep 2026] Title:Generalized Agent Iteration: One Forma

📅 2026-09-15 Vibe Patenting: Evaluating LLM Judges for Professional Patent-Drafting Agents资讯资讯

Vibe Patenting: Evaluating LLM Judges for Professional Patent-Drafting AgentsVibe Patenting: Evaluating LLM Judges for Professional Patent-Drafting Agents

Computer Science > Artificial Intelligence [Submitted on 11 Sep 2026] Title:Vibe Patenting: Evaluating LLM JudgesComputer Science > Artificial Intelligence [Submitted on 11 Sep 2026] Title:Vibe Patenting: Evaluating LLM Judges

📅 2026-09-15 Toward Self-Adaptive Physical AI: Can LLM Agents Manage Long-Horizon Physical Tasks?资讯资讯

Toward Self-Adaptive Physical AI: Can LLM Agents Manage Long-Horizon Physical Tasks?Toward Self-Adaptive Physical AI: Can LLM Agents Manage Long-Horizon Physical Tasks?

Computer Science > Artificial Intelligence [Submitted on 11 Sep 2026] Title:Toward Self-Adaptive Physical AI: CanComputer Science > Artificial Intelligence [Submitted on 11 Sep 2026] Title:Toward Self-Adaptive Physical AI: Can

📅 2026-09-15 LabAgent: Customize Any Research Hubs for Scientific Discoveries Using AI Agents资讯资讯

LabAgent: Customize Any Research Hubs for Scientific Discoveries Using AI AgentsLabAgent: Customize Any Research Hubs for Scientific Discoveries Using AI Agents

Computer Science > Artificial Intelligence [Submitted on 11 Sep 2026] Title:LabAgent: Customize Any Research HubsComputer Science > Artificial Intelligence [Submitted on 11 Sep 2026] Title:LabAgent: Customize Any Research Hubs

📅 2026-09-15 TimeThink: Eliciting Compositional Reasoning in Timeseries Large Language Models资讯资讯

TimeThink: Eliciting Compositional Reasoning in Timeseries Large Language ModelsTimeThink: Eliciting Compositional Reasoning in Timeseries Large Language Models

Computer Science > Artificial Intelligence [Submitted on 11 Sep 2026] Title:TimeThink: Eliciting Compositional ReaComputer Science > Artificial Intelligence [Submitted on 11 Sep 2026] Title:TimeThink: Eliciting Compositional Rea

📅 2026-09-15 Root-Cause Attribution Is a Search Problem: Continual Search for Long-Horizon Agent Failures资讯资讯

Root-Cause Attribution Is a Search Problem: Continual Search for Long-Horizon Agent FailuresRoot-Cause Attribution Is a Search Problem: Continual Search for Long-Horizon Agent Failures

Computer Science > Artificial Intelligence [Submitted on 11 Sep 2026] Title:Root-Cause Attribution Is a Search ProComputer Science > Artificial Intelligence [Submitted on 11 Sep 2026] Title:Root-Cause Attribution Is a Search Pro

📅 2026-09-15 OrchSLM: Probing the Dynamics of Small Language Model Orchestration资讯资讯

OrchSLM: Probing the Dynamics of Small Language Model OrchestrationOrchSLM: Probing the Dynamics of Small Language Model Orchestration

Computer Science > Artificial Intelligence [Submitted on 11 Sep 2026] Title:OrchSLM: Probing the Dynamics of SmallComputer Science > Artificial Intelligence [Submitted on 11 Sep 2026] Title:OrchSLM: Probing the Dynamics of Small

📅 2026-09-15 Token Efficient Task Execution via Application Behavior Modeling for Web Agents资讯资讯

Token Efficient Task Execution via Application Behavior Modeling for Web AgentsToken Efficient Task Execution via Application Behavior Modeling for Web Agents

Computer Science > Artificial Intelligence [Submitted on 11 Sep 2026] Title:Token Efficient Task Execution via AppComputer Science > Artificial Intelligence [Submitted on 11 Sep 2026] Title:Token Efficient Task Execution via App

📅 2026-09-15 Fraglingo: Molecular Design via Attachment-Aware Autoregressive Fragment Generation资讯资讯

Fraglingo: Molecular Design via Attachment-Aware Autoregressive Fragment GenerationFraglingo: Molecular Design via Attachment-Aware Autoregressive Fragment Generation

Computer Science > Artificial Intelligence [Submitted on 11 Sep 2026] Title:Fraglingo: Molecular Design via AttachComputer Science > Artificial Intelligence [Submitted on 11 Sep 2026] Title:Fraglingo: Molecular Design via Attach

📅 2026-09-15 From Legal Text to AI-specific Risk Sources: A Systematic Analysis of the EU AI Act's High-Risk Requirements资讯资讯

From Legal Text to AI-specific Risk Sources: A Systematic Analysis of the EU AI Act's High-Risk RequirementsFrom Legal Text to AI-specific Risk Sources: A Systematic Analysis of the EU AI Act's High-Risk Requirements

Computer Science > Artificial Intelligence [Submitted on 11 Sep 2026] Title:From Legal Text to AI-specific Risk SoComputer Science > Artificial Intelligence [Submitted on 11 Sep 2026] Title:From Legal Text to AI-specific Risk So

📅 2026-09-15 Token Merging for Multilingual Speech Recognition: A Systematic Study Across Model Scale and Fine-Tuning资讯资讯

Token Merging for Multilingual Speech Recognition: A Systematic Study Across Model Scale and Fine-TuningToken Merging for Multilingual Speech Recognition: A Systematic Study Across Model Scale and Fine-Tuning

Computer Science > Computation and Language [Submitted on 7 Jul 2026] Title:Token Merging for Multilingual SpeechComputer Science > Computation and Language [Submitted on 7 Jul 2026] Title:Token Merging for Multilingual Speech

📅 2026-09-15 PhysMent: An Interactive Approach For LLM Reasoning In Physics Problems资讯资讯

PhysMent: An Interactive Approach For LLM Reasoning In Physics ProblemsPhysMent: An Interactive Approach For LLM Reasoning In Physics Problems

Computer Science > Computation and Language [Submitted on 7 Jul 2026] Title:PhysMent: An Interactive Approach ForComputer Science > Computation and Language [Submitted on 7 Jul 2026] Title:PhysMent: An Interactive Approach For

📅 2026-09-15 Lexical Prompt Compression for Large Language Models: A Training-Free, Deterministic Pipeline with Empirical Pareto Analysis Across Eleven Task Categories资讯资讯

Lexical Prompt Compression for Large Language Models: A Training-Free, Deterministic Pipeline with Empirical Pareto Analysis Across Eleven Task CategoriesLexical Prompt Compression for Large Language Models: A Training-Free, Deterministic Pipeline with Empirical Pareto Analysis Across Eleven Task Categories

Computer Science > Computation and Language [Submitted on 8 Jul 2026] Title:Lexical Prompt Compression for Large LComputer Science > Computation and Language [Submitted on 8 Jul 2026] Title:Lexical Prompt Compression for Large L

📅 2026-09-15 TestHallVQA: Exploring LVLMs' Document-Level Reasoning under Redundant Contexts from Scientific Exams资讯资讯

TestHallVQA: Exploring LVLMs' Document-Level Reasoning under Redundant Contexts from Scientific ExamsTestHallVQA: Exploring LVLMs' Document-Level Reasoning under Redundant Contexts from Scientific Exams

Computer Science > Computation and Language [Submitted on 15 Jul 2026] Title:TestHallVQA: Exploring LVLMs' DocumenComputer Science > Computation and Language [Submitted on 15 Jul 2026] Title:TestHallVQA: Exploring LVLMs' Documen

📅 2026-09-15 RFCLLM: Evaluating LLMs' Reasoning Ability of Network Protocol State Machines资讯资讯

RFCLLM: Evaluating LLMs' Reasoning Ability of Network Protocol State MachinesRFCLLM: Evaluating LLMs' Reasoning Ability of Network Protocol State Machines

Computer Science > Computation and Language [Submitted on 11 Sep 2026] Title:RFCLLM: Evaluating LLMs' Reasoning AbComputer Science > Computation and Language [Submitted on 11 Sep 2026] Title:RFCLLM: Evaluating LLMs' Reasoning Ab

📅 2026-09-15 A Hybrid Hierarchical 1D-CNN-BiLSTM Framework for Extractive Summarization of Biomedical and Clinical Text资讯资讯

A Hybrid Hierarchical 1D-CNN-BiLSTM Framework for Extractive Summarization of Biomedical and Clinical TextA Hybrid Hierarchical 1D-CNN-BiLSTM Framework for Extractive Summarization of Biomedical and Clinical Text

Computer Science > Computation and Language [Submitted on 11 Sep 2026] Title:A Hybrid Hierarchical 1D-CNN-BiLSTM FComputer Science > Computation and Language [Submitted on 11 Sep 2026] Title:A Hybrid Hierarchical 1D-CNN-BiLSTM F

📅 2026-09-15 From Token Probabilities to Semantic Constraints: Towards Declarative Probabilistic Evaluation of Language Models资讯资讯

From Token Probabilities to Semantic Constraints: Towards Declarative Probabilistic Evaluation of Language ModelsFrom Token Probabilities to Semantic Constraints: Towards Declarative Probabilistic Evaluation of Language Models

Computer Science > Computation and Language [Submitted on 11 Sep 2026] Title:From Token Probabilities to SemanticComputer Science > Computation and Language [Submitted on 11 Sep 2026] Title:From Token Probabilities to Semantic

📅 2026-09-15 Harmfulness Propagation Dynamics: Layer-wise Trajectories of Adversarial Intent in Large Language Models资讯资讯

Harmfulness Propagation Dynamics: Layer-wise Trajectories of Adversarial Intent in Large Language ModelsHarmfulness Propagation Dynamics: Layer-wise Trajectories of Adversarial Intent in Large Language Models

Computer Science > Computation and Language [Submitted on 11 Sep 2026] Title:Harmfulness Propagation Dynamics: LayComputer Science > Computation and Language [Submitted on 11 Sep 2026] Title:Harmfulness Propagation Dynamics: Lay

📅 2026-09-15 Domain-Specific Jargon in Large Language Models: A Comparative Analysis between General-Purpose and Specialist Models资讯资讯

Domain-Specific Jargon in Large Language Models: A Comparative Analysis between General-Purpose and Specialist ModelsDomain-Specific Jargon in Large Language Models: A Comparative Analysis between General-Purpose and Specialist Models

Computer Science > Computation and Language [Submitted on 11 Sep 2026] Title:Domain-Specific Jargon in Large LanguComputer Science > Computation and Language [Submitted on 11 Sep 2026] Title:Domain-Specific Jargon in Large Langu

📅 2026-09-15 探索RSI，生数新世界模型让机器人开始自我进化资讯资讯

探索RSI，生数新世界模型让机器人开始自我进化中文探索RSI，生数新世界模型让机器人开始自我进化

中文

📅 2026-09-14 Nvidia CEO Jensen Huang tells Trump ‘we’re not going to let [an AI slowdown] happen’资讯资讯

Nvidia CEO Jensen Huang tells Trump ‘we’re not going to let [an AI slowdown] happen’Nvidia CEO Jensen Huang tells Trump ‘we’re not going to let [an AI slowdown] happen’

While Nvidia CEO Jensen Huang was onstage at the All-In Summit on Monday morning in Los Angeles, he got a phone call thaWhile Nvidia CEO Jensen Huang was onstage at the All-In Summit on Monday morning in Los Angeles, he got a phone call tha

📅 2026-09-14 OpenAI buys smartphone camera maker Glass Imaging for $300 million, report says资讯资讯

OpenAI buys smartphone camera maker Glass Imaging for $300 million, report saysOpenAI buys smartphone camera maker Glass Imaging for $300 million, report says

OpenAI has bought smartphone camera maker Glass Imaging in a deal worth over $300 million, according to a report from ThOpenAI has bought smartphone camera maker Glass Imaging in a deal worth over $300 million, according to a report from Th

📅 2026-09-14 Agent Harness vs Agent Framework vs MCP: Which Layer Owns the Loop, State, Tools, Permissions, and Recovery资讯资讯

Agent Harness vs Agent Framework vs MCP: Which Layer Owns the Loop, State, Tools, Permissions, and RecoveryAgent Harness vs Agent Framework vs MCP: Which Layer Owns the Loop, State, Tools, Permissions, and Recovery

Harness, framework, and MCP get used interchangeably in agent architecture discussions. They are not the same thing. TheHarness, framework, and MCP get used interchangeably in agent architecture discussions. They are not the same thing. The

📅 2026-09-14 Reward AI Releases OM-1: A Robot Policy Trained on Human Demonstrations Only, With No Teleoperation or On-Robot Data资讯资讯

Reward AI Releases OM-1: A Robot Policy Trained on Human Demonstrations Only, With No Teleoperation or On-Robot DataReward AI Releases OM-1: A Robot Policy Trained on Human Demonstrations Only, With No Teleoperation or On-Robot Data

Reward AI, a robotics startup whose team’s prior work includes DexCap, HumanPlus, and ALOHA, has released OM-1, short foReward AI, a robotics startup whose team’s prior work includes DexCap, HumanPlus, and ALOHA, has released OM-1, short fo

📅 2026-09-14 Sakana AI Researchers Introduce PC-ALM, a Layer-Local Alternative to Backpropagation That Trains 1000-Layer Networks资讯资讯

Sakana AI Researchers Introduce PC-ALM, a Layer-Local Alternative to Backpropagation That Trains 1000-Layer NetworksSakana AI Researchers Introduce PC-ALM, a Layer-Local Alternative to Backpropagation That Trains 1000-Layer Networks

Backpropagation is a global algorithm: a forward pass, then a backward pass, then a weight update, each locked behind thBackpropagation is a global algorithm: a forward pass, then a backward pass, then a weight update, each locked behind th

📅 2026-09-14 蚂蚁发布大模型内生式安全护栏SingProbe，让AI边生成边识别风险资讯资讯

蚂蚁发布大模型内生式安全护栏SingProbe，让AI边生成边识别风险中文蚂蚁发布大模型内生式安全护栏SingProbe，让AI边生成边识别风险

0 | 本文作者： Nemo | 2026-09-14 21:22 | 大模型正在进入日常问答、办公辅助、客户服务等真实业务场景。用户期待AI快速给出有用的回答，也需要这些回答安全、可靠。如何及时发现不安全内容和编造信息，同时减少安全审核中文0 | 本文作者： Nemo | 2026-09-14 21:22 | 大模型正在进入日常问答、办公辅助、客户服务等真实业务场景。用户期待AI快速给出有用的回答，也需要这些回答安全、可靠。如何及时发现不安全内容和编造信息，同时减少安全审核

📅 2026-09-14 With iOS 27, I’m actually using Siri again资讯资讯

With iOS 27, I’m actually using Siri againWith iOS 27, I’m actually using Siri again

For years, I barely used Siri for anything beyond menial tasks like setting timers or alarms. That has changed with iOSFor years, I barely used Siri for anything beyond menial tasks like setting timers or alarms. That has changed with iOS

📅 2026-09-14 Fashion app Daydream uses Apple Intelligence to help you shop the outfits in your camera roll资讯资讯

Fashion app Daydream uses Apple Intelligence to help you shop the outfits in your camera rollFashion app Daydream uses Apple Intelligence to help you shop the outfits in your camera roll

Daydream, an AI-powered fashion discovery app, launched two new features on Monday to let iPhone users turn saved outfitDaydream, an AI-powered fashion discovery app, launched two new features on Monday to let iPhone users turn saved outfit

📅 2026-09-14 Microsoft’s new AI ‘code of conduct’ tells models not to hack systems or trick humans资讯资讯

Microsoft’s new AI ‘code of conduct’ tells models not to hack systems or trick humansMicrosoft’s new AI ‘code of conduct’ tells models not to hack systems or trick humans

As the AI world shifts its focus to safety and alignment, Microsoft has released a new AI code of conduct meant to guideAs the AI world shifts its focus to safety and alignment, Microsoft has released a new AI code of conduct meant to guide

📅 2026-09-14 Only at TechCrunch Disrupt 2026: What happens when OpenAI ships your roadmap?资讯资讯

Only at TechCrunch Disrupt 2026: What happens when OpenAI ships your roadmap?Only at TechCrunch Disrupt 2026: What happens when OpenAI ships your roadmap?

For AI founders, the biggest competitive threat may no longer be another startup. It may be the platform you’re buildingFor AI founders, the biggest competitive threat may no longer be another startup. It may be the platform you’re building

📅 2026-09-14 Superhuman acquires YC-backed notetaker Fathom as productivity platforms push for agentic work资讯资讯

Superhuman acquires YC-backed notetaker Fathom as productivity platforms push for agentic workSuperhuman acquires YC-backed notetaker Fathom as productivity platforms push for agentic work

Superhuman is joining the cadre of productivity platforms launching notetakers — but rather than building one internallySuperhuman is joining the cadre of productivity platforms launching notetakers — but rather than building one internally

📅 2026-09-14 全球AI大厂集体呼吁“限速” 360：AI安全不能靠企业自审，需第三方攻防把关资讯资讯

全球AI大厂集体呼吁“限速” 360：AI安全不能靠企业自审，需第三方攻防把关中文全球AI大厂集体呼吁“限速” 360：AI安全不能靠企业自审，需第三方攻防把关

0 | 本文作者： Nemo | 2026-09-14 19:43 | 近日，Anthropic、OpenAI、xAI、谷歌DeepMind等全球顶尖AI企业达成罕见行业共识，集体呼吁对前沿AI技术发展实施“全球限速”，通过放缓迭代节奏规中文0 | 本文作者： Nemo | 2026-09-14 19:43 | 近日，Anthropic、OpenAI、xAI、谷歌DeepMind等全球顶尖AI企业达成罕见行业共识，集体呼吁对前沿AI技术发展实施“全球限速”，通过放缓迭代节奏规

📅 2026-09-14 Hear how AI can engineer nature’s comeback at TechCrunch Disrupt 2026资讯资讯

Hear how AI can engineer nature’s comeback at TechCrunch Disrupt 2026Hear how AI can engineer nature’s comeback at TechCrunch Disrupt 2026

Not long ago, bringing an extinct species back to life belonged to science fiction. Today, it’s the mission of a billionNot long ago, bringing an extinct species back to life belonged to science fiction. Today, it’s the mission of a billion

📅 2026-09-14 首届蚂蚁灵波具身大模型挑战赛正式启动资讯资讯

首届蚂蚁灵波具身大模型挑战赛正式启动中文首届蚂蚁灵波具身大模型挑战赛正式启动

首届蚂蚁灵波具身大模型挑战赛正式启动 通过这场大赛，蚂蚁灵波希望将 LingBot-VLA 进一步推向更广泛的开发者社区和高校科研社区 在外滩大会《基座之上：具身智能的场景突围与协同进化》见解论坛上，首届蚂蚁灵波具身大模型挑战赛正式启动。中文首届蚂蚁灵波具身大模型挑战赛正式启动 通过这场大赛，蚂蚁灵波希望将 LingBot-VLA 进一步推向更广泛的开发者社区和高校科研社区 在外滩大会《基座之上：具身智能的场景突围与协同进化》见解论坛上，首届蚂蚁灵波具身大模型挑战赛正式启动。

📅 2026-09-14 全国唯一！商汤大装置临港AIDC获“算效+算电”双5A认证资讯资讯

全国唯一！商汤大装置临港AIDC获“算效+算电”双5A认证中文全国唯一！商汤大装置临港AIDC获“算效+算电”双5A认证

全国唯一！商汤大装置临港AIDC获“算效+算电”双5A认证 业界首个5A算电协同认证 9月12日，在2026中国算力大会“零碳筑基·算电织网——算电协同高质量发展论坛”上，中国信息通信研究院（以下简称“中国信通院”）正式发布算力电力协同测评中文全国唯一！商汤大装置临港AIDC获“算效+算电”双5A认证 业界首个5A算电协同认证 9月12日，在2026中国算力大会“零碳筑基·算电织网——算电协同高质量发展论坛”上，中国信息通信研究院（以下简称“中国信通院”）正式发布算力电力协同测评

📅 2026-09-14 端侧AI从「能跑」到「会进化」，元空智能跑进惠普预装资讯资讯

端侧AI从「能跑」到「会进化」，元空智能跑进惠普预装中文端侧AI从「能跑」到「会进化」，元空智能跑进惠普预装

端侧AI从「能跑」到「会进化」，元空智能跑进惠普预装 让模型和设备共同进化 梦瑶 发自 凹非寺量子位 | 公众号 QbitAI 静态互联网数据的增长红利，见顶了。 下一轮模型能力的跃迁，会越来越依赖与真实环境持续交互的「端侧」。 当整个行业中文端侧AI从「能跑」到「会进化」，元空智能跑进惠普预装 让模型和设备共同进化 梦瑶 发自 凹非寺量子位 | 公众号 QbitAI 静态互联网数据的增长红利，见顶了。 下一轮模型能力的跃迁，会越来越依赖与真实环境持续交互的「端侧」。 当整个行业

📅 2026-09-14 蚂蚁百宝箱推出全新商圈智能体模版，万达率先发布逛街Agent资讯资讯

蚂蚁百宝箱推出全新商圈智能体模版，万达率先发布逛街Agent中文蚂蚁百宝箱推出全新商圈智能体模版，万达率先发布逛街Agent

0 | 本文作者： Nemo | 2026-09-14 15:27 | AI浪潮奔涌而至，线下商圈如何借势破局、实现效率跃升？ 作为企业级智能体开发平台，蚂蚁百宝箱近日正式推出全新的商圈智能体模板，一站式集成美食推荐、店铺导览、营销活动推中文0 | 本文作者： Nemo | 2026-09-14 15:27 | AI浪潮奔涌而至，线下商圈如何借势破局、实现效率跃升？ 作为企业级智能体开发平台，蚂蚁百宝箱近日正式推出全新的商圈智能体模板，一站式集成美食推荐、店铺导览、营销活动推

📅 2026-09-14 NVIDIA Open-Sources OSMO: One YAML Orchestrates Physical AI Training, Simulation, and Robot Testing资讯资讯

NVIDIA Open-Sources OSMO: One YAML Orchestrates Physical AI Training, Simulation, and Robot TestingNVIDIA Open-Sources OSMO: One YAML Orchestrates Physical AI Training, Simulation, and Robot Testing

Robot developers do not have one compute problem. They have 3. A policy is trained on GB200 or H100 clusters, tested inRobot developers do not have one compute problem. They have 3. A policy is trained on GB200 or H100 clusters, tested in

📅 2026-09-14 AI开始改进“改进自己的方法”，RSI进入平方时代丨MetaRSI资讯资讯

AI开始改进“改进自己的方法”，RSI进入平方时代丨MetaRSI中文AI开始改进“改进自己的方法”，RSI进入平方时代丨MetaRSI

AI开始改进“改进自己的方法”，RSI进入平方时代丨MetaRSI 小模型突破瓶颈，旗舰模型打开新空间 允中 发自 凹非寺 量子位 | 公众号 QbitAI 让模型自己生成训练数据、自己改写提示词、自己修复代码，这条路线被称为： 递归自我改中文AI开始改进“改进自己的方法”，RSI进入平方时代丨MetaRSI 小模型突破瓶颈，旗舰模型打开新空间 允中 发自 凹非寺 量子位 | 公众号 QbitAI 让模型自己生成训练数据、自己改写提示词、自己修复代码，这条路线被称为： 递归自我改

📅 2026-09-14 奥特曼被骗！A社一脚油门冲刺IPO，募资叫板SpaceX资讯资讯

奥特曼被骗！A社一脚油门冲刺IPO，募资叫板SpaceX中文奥特曼被骗！A社一脚油门冲刺IPO，募资叫板SpaceX

奥特曼被骗！A社一脚油门冲刺IPO，募资叫板SpaceX 还是太嫩了啊Dario Jay 发自 凹非寺 量子位 | 公众号QbitAI 神了。 呼吁全员踩刹车后，你A社的IPO，还在继续。 彭博社爆料，Anthropic已经选定纳斯达克作为中文奥特曼被骗！A社一脚油门冲刺IPO，募资叫板SpaceX 还是太嫩了啊Dario Jay 发自 凹非寺 量子位 | 公众号QbitAI 神了。 呼吁全员踩刹车后，你A社的IPO，还在继续。 彭博社爆料，Anthropic已经选定纳斯达克作为

📅 2026-09-14 豆包手机助手发布消费者版本 首款新机将于9月16日正式开售资讯资讯

豆包手机助手发布消费者版本 首款新机将于9月16日正式开售中文豆包手机助手发布消费者版本 首款新机将于9月16日正式开售

0 | 本文作者： Nemo | 2026-09-14 10:55 | 9月14日，豆包手机助手消费者版本正式发布。 豆包手机助手以豆包 App 为基础，与手机厂商在系统层面展开合作，围绕交互、任务执行、本地记忆及检索等核心体验，进一步探中文0 | 本文作者： Nemo | 2026-09-14 10:55 | 9月14日，豆包手机助手消费者版本正式发布。 豆包手机助手以豆包 App 为基础，与手机厂商在系统层面展开合作，围绕交互、任务执行、本地记忆及检索等核心体验，进一步探

📅 2026-09-14 TUM 教授 Angela Dai：放下完美数据执念，「逆向自监督」重构 3D 空间智能 | ECCV 2026资讯资讯

TUM 教授 Angela Dai：放下完美数据执念，「逆向自监督」重构 3D 空间智能 | ECCV 2026中文TUM 教授 Angela Dai：放下完美数据执念，「逆向自监督」重构 3D 空间智能 | ECCV 2026

2D去噪范式在真实3D遮挡前失效，破局关键在于将物理机制转化为结构先验。    作者丨张璐    编辑丨岑峰            &nbs中文2D去噪范式在真实3D遮挡前失效，破局关键在于将物理机制转化为结构先验。    作者丨张璐    编辑丨岑峰            &nbs

📅 2026-09-14 R2VC: Modular Fact-Checking with Retrieval, Verification, and Confidence Calibration资讯资讯

R2VC: Modular Fact-Checking with Retrieval, Verification, and Confidence CalibrationR2VC: Modular Fact-Checking with Retrieval, Verification, and Confidence Calibration

Computer Science > Computation and Language [Submitted on 6 Aug 2026] Title:R2VC: Modular Fact-Checking with RetriComputer Science > Computation and Language [Submitted on 6 Aug 2026] Title:R2VC: Modular Fact-Checking with Retri

📅 2026-09-14 The Cost of Compression: A Rate-Distortion Limit on Factual Hallucination资讯资讯

The Cost of Compression: A Rate-Distortion Limit on Factual HallucinationThe Cost of Compression: A Rate-Distortion Limit on Factual Hallucination

Computer Science > Computation and Language [Submitted on 10 Sep 2026] Title:The Cost of Compression: A Rate-DistoComputer Science > Computation and Language [Submitted on 10 Sep 2026] Title:The Cost of Compression: A Rate-Disto

📅 2026-09-14 Local Edits, Global Ripples: Replay-Informed Policy Adaptation for Workflow Synthesis资讯资讯

Local Edits, Global Ripples: Replay-Informed Policy Adaptation for Workflow SynthesisLocal Edits, Global Ripples: Replay-Informed Policy Adaptation for Workflow Synthesis

Computer Science > Computation and Language [Submitted on 10 Sep 2026] Title:Local Edits, Global Ripples: Replay-IComputer Science > Computation and Language [Submitted on 10 Sep 2026] Title:Local Edits, Global Ripples: Replay-I

📅 2026-09-14 GAUGE: When Not to Trust LLM-as-a-Judge in User-Simulated Evaluation of Task-Oriented Agents资讯资讯

GAUGE: When Not to Trust LLM-as-a-Judge in User-Simulated Evaluation of Task-Oriented AgentsGAUGE: When Not to Trust LLM-as-a-Judge in User-Simulated Evaluation of Task-Oriented Agents

Computer Science > Computation and Language [Submitted on 10 Sep 2026] Title:GAUGE: When Not to Trust LLM-as-a-JudComputer Science > Computation and Language [Submitted on 10 Sep 2026] Title:GAUGE: When Not to Trust LLM-as-a-Jud

📅 2026-09-14 Repair Before Reinforce: Context-Augmented Knowledge Graph Reasoning for Multi-Hop Question Answering资讯资讯

Repair Before Reinforce: Context-Augmented Knowledge Graph Reasoning for Multi-Hop Question AnsweringRepair Before Reinforce: Context-Augmented Knowledge Graph Reasoning for Multi-Hop Question Answering

Computer Science > Computation and Language [Submitted on 10 Sep 2026] Title:Repair Before Reinforce: Context-AugmComputer Science > Computation and Language [Submitted on 10 Sep 2026] Title:Repair Before Reinforce: Context-Augm

📅 2026-09-14 Chopthin-Consensus Power Sampling: A Diversity-Preserving Approach to LLM Decoding资讯资讯

Chopthin-Consensus Power Sampling: A Diversity-Preserving Approach to LLM DecodingChopthin-Consensus Power Sampling: A Diversity-Preserving Approach to LLM Decoding

Computer Science > Computation and Language [Submitted on 10 Sep 2026] Title:Chopthin-Consensus Power Sampling: AComputer Science > Computation and Language [Submitted on 10 Sep 2026] Title:Chopthin-Consensus Power Sampling: A

📅 2026-09-14 EAR: Entity-Aware Partitioning Approach for Retrieval-Augmented Generation Development资讯资讯

EAR: Entity-Aware Partitioning Approach for Retrieval-Augmented Generation DevelopmentEAR: Entity-Aware Partitioning Approach for Retrieval-Augmented Generation Development

Computer Science > Computation and Language [Submitted on 10 Sep 2026] Title:EAR: Entity-Aware Partitioning ApproaComputer Science > Computation and Language [Submitted on 10 Sep 2026] Title:EAR: Entity-Aware Partitioning Approa

📅 2026-09-14 Breaking the Token Ceiling: Distilling Smaller, Stronger Byte Models资讯资讯

Breaking the Token Ceiling: Distilling Smaller, Stronger Byte ModelsBreaking the Token Ceiling: Distilling Smaller, Stronger Byte Models

Computer Science > Computation and Language [Submitted on 11 Sep 2026] Title:Breaking the Token Ceiling: DistillinComputer Science > Computation and Language [Submitted on 11 Sep 2026] Title:Breaking the Token Ceiling: Distillin

📅 2026-09-14 ESTS at WMT26: Routing-Informed Expert Pruning for Model Compression资讯资讯

ESTS at WMT26: Routing-Informed Expert Pruning for Model CompressionESTS at WMT26: Routing-Informed Expert Pruning for Model Compression

Computer Science > Computation and Language [Submitted on 11 Sep 2026] Title:ESTS at WMT26: Routing-Informed ExperComputer Science > Computation and Language [Submitted on 11 Sep 2026] Title:ESTS at WMT26: Routing-Informed Exper

📅 2026-09-14 智谱提前剧透GLM-6.0：完全自训练方法公开了资讯资讯

智谱提前剧透GLM-6.0：完全自训练方法公开了中文智谱提前剧透GLM-6.0：完全自训练方法公开了

智谱提前剧透GLM-6.0：完全自训练方法公开了 模型、论文都还没出，没想到先从财务公告上剧透了 梦晨 发自 凹非寺 量子位 | 公众号 QbitAI 上市不到九个月，智谱一口气再融335亿元。 智谱在港交所发布公告：配售新H股加发行201中文智谱提前剧透GLM-6.0：完全自训练方法公开了 模型、论文都还没出，没想到先从财务公告上剧透了 梦晨 发自 凹非寺 量子位 | 公众号 QbitAI 上市不到九个月，智谱一口气再融335亿元。 智谱在港交所发布公告：配售新H股加发行201

📅 2026-09-14 Anthropic’s 3-Step ‘Pace the Frontier’ Plan Wins OpenAI, xAI and Microsoft Support: Is It Too Late to Slow AI Down?资讯资讯

Anthropic’s 3-Step ‘Pace the Frontier’ Plan Wins OpenAI, xAI and Microsoft Support: Is It Too Late to Slow AI Down?Anthropic’s 3-Step ‘Pace the Frontier’ Plan Wins OpenAI, xAI and Microsoft Support: Is It Too Late to Slow AI Down?

On September 12, 2026, Anthropic CEO Dario Amodei published a writeup ‘We Must Pace the Frontier’. Its core message is bOn September 12, 2026, Anthropic CEO Dario Amodei published a writeup ‘We Must Pace the Frontier’. Its core message is b

📅 2026-09-14 资讯

OpenAI 发布 GPT-6 Astra：「欢迎来到 AGI 时代」，API 定价 $10/$50 为上代 2.5 倍

AI 导航讯——OpenAI 于 2026-09-03 发布新一代旗舰 GPT-6 Astra：ARC-AGI-3 99.9%、ExploitBench 100%，首个达「关键」网络安全等级的 OpenAI 模型；API 定价输入 $10、

📅 2026-09-03 资讯

Google 三周内再推 Gemini 3.8 Flash：介绍价 $0.75/$3.75，Cyber 网络安全变体齐发

AI 导航讯——Google 于 2026-09-02 发布 Gemini 3.8 Flash 与 3.8 Flash Cyber：介绍价输入 $0.75、输出 $3.75 每百万 tokens（2027 年起翻倍），DeepSWE 73.

📅 2026-09-03 资讯

WorkBuddy 开放平台解析：用 Skill、Connector、专家与知识库给 AI 同事「配工具」

AI 导航讯——WorkBuddy 把 AI 办公工作台做成可扩展的开放平台（扩展体系）：通过 Skill、Connector/MCP、专家、企业知识库四大机制扩展能力，配合 Open Platform Token 安全访问用户资源与 RR

📅 2026-09-03 资讯

微信支付 AI 专属卡上新：一张卡直连 DeepSeek Harness 与 OpenClaw

AI 导航讯——微信支付推出 AI 专属卡，面向 AI 开发者与重度使用者，支持把卡片额度接入 DeepSeek Harness 与 OpenClaw 两类 Agent 工具，微信内一键开通、统一结算 AI 支出，进一步降低国内智能体开发门

📅 2026-08-31 资讯

Anthropic 发布 MHS 模型硬件标准：让 AI 像插 USB-C 一样操控真实世界设备

AI 导航发布 Anthropic MHS（Model Hardware Standard）模型硬件标准技术解读：2026-08-27 发布的研究预览，用统一驱动规范让任意大模型智能体安全操控显微镜、机械臂、量子激光等物理设备，被称为「AI

📅 2026-08-28
返回首页